1. A deformation resistance prediction system, comprising: a rolling condition determination device configured to determine a rolling condition to be set in a rolling device; a storage device configured to collect operation data of rolling performed by the rolling device and the rolling condition; and a deformation resistance prediction device including a processor and a memory and configured to predict a deformation resistance of a rolled material based on the operation data and the rolling condition, wherein the deformation resistance prediction device includes a rolling data acquisition unit configured to acquire the operation data and the rolling condition, accumulate the operation data and the rolling condition as rolling performance data, and acquire a rolling condition candidate from the rolling condition determination device, a prediction model configured to predict the deformation resistance of the rolled material, and a deformation resistance prediction unit configured to calculate a prediction value of the deformation resistance of the rolled material using the prediction model, the prediction model includes a global model determined based on the entire rolling performance data and set in advance as a prediction model for estimating a deformation resistance of each steel type of the rolled material, and a class model determined based on rolling performance data of a deformation resistance of a specific steel type in which alloy components of the rolled material are similar or common among the rolling performance data and set in advance as the prediction model for estimating the deformation resistance of the specific steel type, and the deformation resistance prediction unit calculates the prediction value of the deformation resistance of the rolled material based on the rolling condition candidate using the global model and the class model.

2. The deformation resistance prediction system according to claim 1, wherein the deformation resistance prediction unit includes a global model application unit configured to apply the global model to the rolling condition candidate to calculate a global prediction value of the deformation resistance of the rolled material under the rolling condition candidate, and a class model application unit configured to apply the class model to the rolling condition candidate, calculate a class correction value of the deformation resistance of the rolled material under the rolling condition candidate, and calculate a prediction value of the deformation resistance under the rolling condition candidate based on the class correction value and the global prediction value.

3. The deformation resistance prediction system according to claim 1, wherein the deformation resistance prediction unit further includes an accuracy evaluation unit configured to calculate a prediction accuracy of the prediction model based on the prediction value of the deformation resistance and an actual value of the deformation resistance included in the rolling performance data.

4. The deformation resistance prediction system according to claim 3, further comprising: an accuracy display unit configured to generate a change history of a prediction accuracy of the class model based on the prediction accuracy of the prediction model.

5. The deformation resistance prediction system according to claim 1, wherein the deformation resistance prediction unit further includes a class model adjustment unit configured to adjust the class model based on an actual value of the deformation resistance of the specific steel type included in the rolling performance data.

6. The deformation resistance prediction system according to claim 5, wherein the class model adjustment unit adjusts the class model using the rolling performance data of the specific steel type for the selected class model.

7. The deformation resistance prediction system according to claim 3, wherein the deformation resistance prediction unit further includes an abnormal data exclusion unit configured to exclude, as abnormal data, data in which a value of the rolling performance data exceeds a predetermined range, and the accuracy evaluation unit calculates the prediction accuracy of the prediction model based on the rolling performance data excluding the abnormal data.

8. The deformation resistance prediction system according to claim 1, wherein the global model is configured by multiplying two factors, a first factor of the two factors being configured to include an addition result of a constant term, a logarithmic term of a strain rate of the rolling, and a power term of a temperature, and a second factor of the two factors being configured to include a power of a strain of the rolling.

9. The deformation resistance prediction system according to claim 1, wherein the class model is configured by multiplying two factors, a first factor of the two factors being configured by an exponential function having an exponential part including an addition result of a constant term, a logarithmic term of a strain rate of the rolling, and a power term of a temperature, and a second factor of the two factors being configured to include a power of a strain of the rolling.

10. A deformation resistance prediction method in which a deformation resistance prediction device including a processor and a memory predicts a deformation resistance of a rolled material, the method comprising: a first step in which the deformation resistance prediction device acquires a rolling condition candidate from a rolling condition determination device configured to determine a rolling condition to be set in a rolling device; a second step in which the deformation resistance prediction device acquires, from a storage device configured to collect operation data of rolling performed by the rolling device and the rolling condition, the operation data and the rolling condition, and accumulates the operation data and the rolling condition as rolling performance data; and a third step in which the deformation resistance prediction device calculates a prediction value of the deformation resistance of the rolled material using a prediction model for predicting the deformation resistance of the rolled material, wherein in the third step, the prediction value of the deformation resistance of the rolled material is calculated based on the rolling condition candidate using the prediction model including a global model determined based on the entire rolling performance data and set in advance as a model for estimating a deformation resistance of each steel type of the rolled material, and a class model determined based on rolling performance data of a deformation resistance of a specific steel type in which alloy components of the rolled material are similar or common among the rolling performance data and set in advance as a model for estimating the deformation resistance of the specific steel type.

11. A deformation resistance prediction device that includes a processor and a memory and configured to predict a deformation resistance of a rolled material, the deformation resistance prediction device comprising: a rolling data acquisition unit configured to acquire operation data of rolling and a rolling condition, accumulate the operation data and the rolling condition as rolling performance data, and receive a rolling condition candidate; a prediction model configured to predict a deformation resistance of the rolled material; and a deformation resistance prediction unit configured to calculate a prediction value of the deformation resistance of the rolled material using the prediction model, wherein the prediction model includes a global model determined based on the entire rolling performance data and set in advance as a prediction model for estimating a deformation resistance of each steel type of the rolled material, and a class model determined based on rolling performance data of a deformation resistance of a specific steel type in which alloy components of the rolled material are similar or common among the rolling performance data and set in advance as the prediction model for estimating the deformation resistance of the specific steel type, and the deformation resistance prediction unit calculates the prediction value of the deformation resistance of the rolled material based on the rolling condition candidate using the global model and the class model.

DEFORMATION RESISTANCE PREDICTION SYSTEM, DEFORMATION RESISTANCE PREDICTION METHOD, AND DEFORMATION RESISTANCE PREDICTION DEVICE CLAIM OF PRIORITY The present application claims priority from Japanese patent application JP 2021-055335 filed on March 29, 2021, the content of which is hereby incorporated by reference into this application. BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a deformation resistance prediction system, a deformation resistance prediction method, and a deformation resistance prediction device in a hot rolling line. 2. Description of the Related Art In hot rolling, in order to form a slab (steel piece) having a thickness of several tens of centimeters into a rolled material having a thickness of 1 mm to 10 mm, the slab is heated to about 1200°C and then rolled a plurality of times. Rolling at a time is generally referred to as a "pass". In order to produce a rolled material having a desired size and shape from a slab, it is necessary to determine a rolling amount for each rolling pass. Therefore, in a hot rolling line control system, a mean flow stress generated by rolling, that is, a deformation resistance is calculated in advance, and the rolling amount for each rolling pass is determined by using the calculated amount. When an error of a calculation result of the deformation resistance is large, setting of the rolling amount becomes inappropriate, and a rolling failure occurs. As a part of the hot rolling line control system, a device or a subsystem for predicting the deformation resistance is a deformation resistance prediction device, and a method in which the deformation resistance prediction device calculates the deformation resistance based on information input from a higher-level hot rolling line control system is a deformation resistance prediction method. As a technique for predicting the deformation resistance, a technique disclosed in Theoretical and Practice of Plate-Rolling, edited by The Iron and Steel Institute of Japan, p. 161 (1992) (NPL 1) is known. According to this technique, a deformation resistance Kp is calculated based on a carbon concentration [C] and an absolute temperature T of the rolled material, a strain e and a strain rate sdot at the time of rolling as in Equation (1) using the following mathematical model. The mathematical model is a model obtained by collectively deriving results of experiments using many combinations of [C], T, s, and sdot. Kp = 9.8 exp{ 0.126 -1.75[C] + 0.94[C]2) x Exp{ (2851 + 2968[C] -1120[C]2) / T) x eO.21 x £dot°12 As another technique for predicting the deformation resistance, a technique in V.B. Ginzburg, Metallurgical Design of Flat Rolled Steels, p. 295 (2005) (NPL 2) is known. In this technique, as in Equation (2), the deformation resistance Kp is calculated based on a rotation speed V of a rolling roll, a rolling reduction r, a plate thickness h after rolling, a temperature T of the rolled material, a radius R of a work roll, and a rolling reduction amount A using the following mathematical model. In this mathematical model, a0 to a8 are coefficients determined for each steel type using rolling performance data. Kp = (a0 + a^ + a2V2 +a3r + a4r2 + a5h + a6h2 + a7T + a8T2)/(RA)1/2 2 As yet another technique for predicting the deformation resistance, a technique in JP-A-H10-109106 (PTL 1) is known. In this technique, component membership functions Fi(C), F2(C), and F3(C) using contents of chemical components as parameters are set respectively for a plurality of specific steel types Ai, A2, and A3 for which deformation resistances are to be calculated in advance, deformation resistances of steel types Bi, B2, B3, and B4 other than the specific steel types are calculated by calculating a function value corresponding to the content of the chemical component based on the above membership functions, and calculating matching degrees of the specific steel types, and the deformation resistance is calculated based on the matching degrees and the deformation resistances of the specific steel types. According to PTL 1, this technique makes it possible to accurately calculate a deformation resistance of a steel type other than specific steel types by a relatively simple procedure. As still another technique for predicting the deformation resistance, a technique in JP-A-2010-207900 (PTL 2) is known. In this technique, a database creation step of storing past performance data as a database by using factors inflecting the deformation resistance in hot rolling and the deformation resistance as an explanatory variable and an objective variable, respectively, a request point data input step of inputting, as request point data, data of the explanatory variable corresponding to the deformation resistance to be predicted, a neighborhood data selecting step of calculating a distance between data stored in the database and the request point data, and selecting data having a short calculated distance as neighborhood data, and a local model creation step of creating a local model for locally fitting a neighborhood of a request point based on the selected neighborhood data are performed, and the deformation resistance is predicted based on the created local model and the request point data. According to PTL 2, with this technique, it is possible to perform deformation resistance prediction with higher accuracy than that in the related art, and it is not necessary to create a global approximate expression, and thus it is possible to save time and effort such as maintenance at the time of application. However, in the technique disclosed in NPL 1, since it is very difficult to create the mathematical model even if it is not impossible to create a mathematical model covering a wide variety of rolled materials and rolling conditions based on experimental results, in a rolled material having an alloy composition different from that of the experimental material, a prediction error of the deformation resistance is large, and there remains a possibility that a rolling failure occurs. In the technique disclosed in NPL 2, since coefficients of the mathematical model are determined for each steel type using the rolling performance data, it is difficult to accurately determine the coefficients of the mathematical model for a new steel type having no performance data, a prediction error of the deformation resistance with respect to a new steel type is large, and there is a possibility that a rolling failure occurs. In the technique disclosed in PTL 1, since the deformation resistance is predicted by using the matching degree calculated based on the membership functions Fi (C) , F2 (C) , and F3 (C) for each alloy component for each of the steel types Ai, A2, and A3 for which the deformation resistance has been calculated in advance, when the membership function is changed or the steel type for which the deformation resistance has been calculated in advance is changed, a prediction result of the deformation resistance is changed for all the rolled materials. Therefore, even if there is a steel type having low a prediction accuracy of the deformation resistance among various steel types, it is difficult to improve the prediction accuracy by adjusting a mathematical model limited to the steel type, and in such a steel type, a prediction error of the deformation resistance is large, and there is a possibility that a rolling failure occurs. In the technique disclosed in PTL 2, since the local model in which only data close to the request point data is fitted is used, it is difficult to sufficiently utilize a tendency indicated by a large number of pieces of data in the model, in the case of a steel type having a small amount of neighborhood data, when an error is included in the neighborhood data for some reason, the prediction error of the deformation resistance is large due to the influence of the error, and there remains a possibility that a rolling failure occurs. SUMMARY OF THE INVENTION In view of the above, an object of the invention is to provide a deformation resistance prediction system, a deformation resistance prediction method, and a deformation resistance prediction device capable of accurately predicting the deformation resistance even for a steel type having a relatively small amount of rolling performance data and capable of easily making adjustment for each steel type. In order to solve the above problems, the invention provides a deformation resistance prediction system, including: a rolling condition determination device configured to determine a rolling condition to be set in a rolling device; a storage device configured to collect operation data of rolling performed by the rolling device and a rolling condition; and a deformation resistance prediction device including a processor and a memory and configured to predict a deformation resistance of a rolled material based on the operation data and the rolling condition, in which the deformation resistance prediction device includes a rolling data acquisition unit configured to acquire the operation data and the rolling condition, accumulate the operation data and the rolling condition as rolling performance data, and acquire a rolling condition candidate from the rolling condition determination device, a prediction model configured to predict the deformation resistance of the rolled material, and a deformation resistance prediction unit configured to calculate a prediction value of the deformation resistance of the rolled material using the prediction model, the prediction model includes a global model determined based on the entire rolling performance data and set in advance as a prediction model for estimating a deformation resistance of each steel type of the rolled material, and a class model determined based on rolling performance data of a deformation resistance of a specific steel type in which alloy components of the rolled material are similar or common among the rolling performance data and set in advance as the prediction model for estimating the deformation resistance of the specific steel type, and the deformation resistance prediction unit calculates the prediction value of the deformation resistance of the rolled material based on the rolling condition candidate using the global model and the class model. By using the technique of the invention, the deformation resistance is predicted by overlapping the global model of the deformation resistance determined by using the rolling performance data of a large number of steel types and the class model of the deformation resistance determined by using the rolling performance data of one or a small number of steel types, and thereby even for a steel type having a relatively small amount of rolling performance data, it is possible to predict the deformation resistance at least with an accuracy equal to or higher than an accuracy of the global model of the deformation resistance. In addition, since the class model is selected and adjusted, it is easy to adjust only a class model requiring adjustment. Further, since the method includes a step of evaluating the prediction accuracy of each class model using the rolling performance data and a step of displaying a change history of the prediction accuracy of each class model, it is easy for the user to select a class model requiring adjustment. BRIEF DESCRIPTION OF THE DRAWINGS FIG. 1 is a configuration diagram showing an example of a schematic configuration of a hot rolling line control system according to an embodiment of the invention, which shows a configuration of the hot rolling line control system according to the invention. FIG. 2 is a flowchart showing an example of a step performed by a deformation resistance prediction device of the invention. FIG. 3 is a flowchart showing an example of a deformation resistance prediction step performed by the deformation resistance prediction device of the invention. FIG. 4 is a diagram showing an example of a rolling condition of the invention. FIG. 5 is a diagram showing an example of a global model stored in the deformation resistance prediction device of the invention. FIG. 6 is a flowchart showing an example of a prediction accuracy evaluation step performed by the deformation resistance prediction device of the invention. FIG. 7 is a diagram showing an example of a screen displayed by an accuracy display step performed by the deformation resistance prediction device of the invention. FIG. 8 is a diagram showing another example of the screen displayed by the accuracy display step performed by the deformation resistance prediction device of the invention. FIG. 9 is a diagram showing an example of a screen on which the deformation resistance prediction device of the invention determines a threshold value and determines whether model adjustment is necessary. FIG. 10 is a diagram showing an example of a screen on which data for determining that model adjustment is unnecessary by the deformation resistance prediction device of the invention is displayed. FIG. 11 is a flowchart showing an example of a class model adjustment step performed by the deformation resistance prediction device of the invention. FIG. 12 is a diagram showing an example of a class model stored in the deformation resistance prediction device of the invention. FIG. 13 is a distribution diagram showing an example of comparison between actual values and prediction values of a deformation resistance according to the deformation resistance prediction method of the invention. FIG. 14 is a distribution diagram showing an example of comparison between actual values and prediction values of a deformation resistance in the related art. FIG. 15 is a configuration diagram showing an example of a schematic configuration of the deformation resistance prediction device according to the invention. DESCRIPTION OF THE PREFERRED EMBODIMENTS Hereinafter, embodiments of the invention will be described in detail with reference to the drawings. In the drawings, the same elements are denoted by the same reference numerals, and redundant descriptions thereof are omitted. FIG. 1 shows a hot rolling line control system 1 (or a deformation resistance prediction system) according to the invention. The hot rolling line control system 1 determines a rolling condition 110 based on higher-level system information input from a higher-level system 2, and sets the rolling condition 110 in a hot rolling line 3 (rolling device) . The hot rolling line control system 1 includes a rolling condition determination device 11, a deformation resistance prediction device 12, a rolling performance data storage device 13, a rolling condition setting device 14, and an operation data collection device 15. The rolling condition determination device 11 determines the rolling condition 110 based on the information of the higher-level system 2. In order to determine the rolling condition 110, the rolling condition determination device 11 calculates prediction values of deformation resistances in various rolling conditions 110 by using the deformation resistance prediction device 12, and determines an optimum rolling condition 110 within a range satisfying a constraint condition caused by equipment conditions and the like of the hot rolling line 3. The rolling condition determination device 11 stores the determined rolling condition 110 in the rolling performance data storage device 13, and then outputs the rolling condition 110 to the rolling condition setting device 14. The rolling condition setting device 14 sets the rolling condition 110 determined by the rolling condition determination device 11 in the hot rolling line 3. Various sensors (for example, temperature, position, or thickness) are arranged in the rolling line 3, and operation data 150 is measured during an operation of the rolling line 3. The operation data 150 is collected by the operation data collection device 15 and stored in the rolling performance data storage device 13. The deformation resistance prediction device 12 performs prediction accuracy evaluation of a prediction model 123 (mathematical model) used for prediction of the deformation resistance and coefficient adjustment of a mathematical model using rolling performance data 130 generated together with the rolling condition 110 and the operation data 150 stored in the rolling performance data storage device 13 and a rolling condition candidate 113 from the rolling condition determination device 11. The deformation resistance prediction device 12 predicts the deformation resistance when receiving the rolling condition candidate 113 from the rolling condition determination device 11. The rolling condition candidate 113 is constituted by data similar to that of the rolling condition 110, and includes a steel type of a rolled material to be rolled from now on and a set value (expected value) of rolling. The prediction model 123 includes a global model 127 and a class model 128 set in advance. The class model 128 is appropriately adjusted in a class model adjustment step as described later. The global model 127 predicts a general deformation resistance of steel types using a prediction model determined in advance by using the rolling performance data 130 of a large number (for example, several tens of types) of steel types. In other words, the global model 127 is a model for predicting a tendency of the deformation resistance according to the steel type based on the already collected rolling performance data 130. On the other hand, the class model 128 is a prediction model determined in advance using rolling performance data of one or a predetermined number (for example, five types) of steel types (classes) having similar or common alloy components among the collected rolling performance data 130. In other words, the global model 127 can be a set of a plurality of class models 128. As will be described later, the deformation resistance prediction device 12 captures, with the global model 127, a general (global) tendency of the deformation resistance of the rolled material, captures, with the class model 128, a tendency of the deformation resistance in a class different from the general tendency, calculates a deformation resistance prediction value Kp, and provides the deformation resistance prediction value Kp to the rolling condition determination device 11. A configuration of the deformation resistance prediction device 12 of the invention will be described after describing a series of steps performed by the deformation resistance prediction device 12 . FIG. 2 shows an example of processing performed by the deformation resistance prediction device 12 of the invention. When the rolling condition is set, the deformation resistance prediction device 12 performs a deformation resistance prediction step SI (deformation resistance prediction unit). This processing can be executed when an administrator or the like of the hot rolling line control system 1 sets a new rolling condition 110 based on the information of the higher-level system 2 and the rolling condition candidate 113 is output from the rolling condition determination device 11 to the deformation resistance prediction device 12. When the rolling performance data 130 is stored in the deformation resistance prediction device 12, a prediction accuracy evaluation step S2 (prediction accuracy evaluation unit) is performed. Based on the evaluated prediction accuracy, an adjustment necessity determination step S3 is performed for the prediction model 123. When it is determined that adjustment of the prediction model 123 is necessary, a class model adjustment step S4 is performed. Hereinafter, each step of the present embodiment will be disclosed in detail. FIG. 3 shows detailed steps of the deformation resistance prediction step SI performed by the deformation resistance prediction device 12 of the invention. When the deformation resistance prediction step SI is started, the deformation resistance prediction device 12 performs a rolling condition acquisition step S101 to acquire the rolling condition candidate 113 from the rolling condition determination device 11. Next, the deformation resistance prediction device 12 performs a global model application step S102 to apply the global model 127 of the deformation resistance stored in the deformation resistance prediction device 12 to the rolling condition candidate 113, and to calculate a global prediction value KPfG of the deformation resistance. The global prediction value KPfG is a prediction value indicating a general tendency of the deformation resistances of the steel types. Next, the deformation resistance prediction device 12 performs a class model application step S103 to apply the class model 128 of the deformation resistance to the rolling condition candidate 113, to calculate a class correction value APfC of the deformation resistance, and to calculate the prediction value Kp of the deformation resistance based on the class correction value APfC and the global prediction value KPfG of the deformation resistance. The class correction value APfC is a correction value for obtaining the prediction value of a deformation resistance unique to the alloy component of the corresponding steel type. In the invention, although a class of the rolled material is a unit in which rolled materials are classified based on a similarity of a chemical composition (alloy component), and typically the steel types are the same, a plurality of steel types maybe set as one class, or conversely, one steel type may be divided into a plurality of classes. The setting of the class may be defined by an administrator or a user of the deformation resistance prediction device 12. In the invention, by using both the global model 127 of the deformation resistance and the class model 128 of the deformation resistance in this manner, a global tendency indicated by data of a plurality of classes is captured by the global model 127 of the deformation resistance, and a tendency of data in a class different from the global tendency is captured by the class model 128. Next, the deformation resistance prediction device 12 performs a deformation resistance prediction value output step S104 to output the deformation resistance prediction value to the rolling condition determination device 11, and then ends the deformation resistance prediction step SI. One simple embodiment of the global model 127 and the class model 128 of the deformation resistance in the invention will be described below. One embodiment of a method of predicting the deformation resistance by using overlapping of the global model 127 and the class model 128 is an example in which the prediction value Kp of the deformation resistance is represented by a product of the global prediction value KP,G and the class correction value Ap,c of the deformation resistance. Kp = KPQ x Apc (3) Another embodiment of the method of predicting the deformation resistance by using overlapping of the global model 127 and the class model 128 is an example in which the prediction value Kp of the deformation resistance is represented by a sum of the global prediction value KP,G and the class correction value Ap,c of the deformation resistance. Kp - Kp G + Ap c (4) The global model 127 of the deformation resistance is in a form in which two factors are multiplied, and one factor is configured in a form including addition of a constant term, a logarithmic term of a strain rate, and an exponentiation term of a temperature, and the other factor is configured in a form including a power of a strain. An example of such a global model 127 is shown in the following equation (5). KD G = [ G0 + Gi In(edot) + G2/T ] x t^ KPG = [ Go + Gi In(edot) + G2/T + G3XT]XEG4 Kp^o = [ G0 + G-, In(edot) + G2/T + G3/T2 ] x £G4 In examples of the global model 127 described above, Go, Gi, G2, G3, G4, and the like are functions of alloy components of a rolled material, and can be expressed by a mathematical model shown in the following equation (6) as one simple embodiment. G0 = G0]Const + G0]C Wc + G0,si WSi + G0]Mn WMn + VJ-J — v-»i const + G1]CWc+G1]SiWSi+G1]MnWMn + vj2 ~ v-'2,const T G2]C Wc + G2iSi WSj + G2]Mn WMn + ^3 ~ VJ4Si WSim + ... + Gi E!m_N WE|m_Nm)ln(edotm) + (^const + ^2,0 WCm + G2si WSi m + ... + G2j E|m_N WE|m_N m)/Tm (8) All M pieces of performance data can be represented by the same expression. When the M expressions are expressed in a form of a matrix, the following expression is obtained. Y = LXX (9) Here, Y on the left side of Equation (9) is a vector of M rows, and the mth row thereof is represented by the following equation. Y = K / p G3,const n m 1 m rvp,act,m ' cm \±u> In the above equation (9), L on the right side is a matrix of M rows and 3N + 3 columns, and the mth row thereof is given by the following equation. Lm>{i!...JN+i} _ 1. WCm, WSim,..., WE|m_Nm U = ln(£dotm), ln(£dotm)WCjm, ln(£dotm)WSim, .... ln(£dotm) WE|mNm LmJ{2N+3J...,3N+3} = ^m* WCm/Tm, WSi m/Tm, ..., WEim_N m /Tm (n) In the above equation (9), X on the right side is a vector of 3N + 3 rows including parameters of the mathematical model of the global coefficient, and an element thereof is calculated by the following equation (12) . X^.^N+I} = G0const' G0>C, G(),Si> •■•, G(),E!m_N X{N+2,...,2N+2} = GljConst> ^1,C> ^1sSh •••> Ql,E!m_N X{2N+3,...,3N+3} = G2,const. G2>c> G2si> •■-, G2,Elm_N (12 Therefore, parameters of the mathematical model of the global coefficient are obtained by calculating the vector X at which a square error |Y-LxX|2 is minimized for each of G3,COnst while changing G3,COnst, and selecting G3,COnst and X at which the square error is further minimized. A method of calculating the vector X at which the square error |Y-LxX|2 is minimized is well known in the field of scientific and technical calculation, and the square error can be easily calculated by using a numerical analysis library of an open source such as a linear algebra package (LAPCK) , for example. Although in the above description, all of the Go, Gi, and G2 include the terms of N types of alloy components, and G3 includes only the constant term, even when the terms of the alloy components included in G0, Gi, G2, G3, and G4 are different from each other, the parameters can be determined using the performance data in a similar manner as described above. Although in the above description, an example in which the mathematical model of the global coefficient is a linear function of the alloy component has been described, the mathematical model of the global coefficient may be a nonlinear function. In the case of a nonlinear function, the parameter can be determined using a nonlinear optimization method of a continuous function, and as such a nonlinear optimization method, many methods such as a Newton method, a simplex method, a simulation annealing method, and a Levenberg-Marquardt method are well known. As shown in the above example, the parameters of the global model 127 of the deformation resistance of the invention are determined using the rolling performance data 130 of a large number of steel types. The class model 128 of the deformation resistance is obtained by multiplying two factors, and one factor is configured by an exponential function having an exponential part including a sum of an exponential term, a logarithmic term of a strain rate, and an exponentiation term of a temperature, and one factor is configured in a form including a power of a strain. An example of such a class model 128 is shown in the following equation (13) . ApC = exp( C0 + C1 In(edot) + C2/T )x £C3 Ap c = exp( C0 + C1 In(edot) + C2/T + C3 x T) x £C4 Ap'c = exp( C0 + C1 In(edot) + C2/T + C3/T2) x £C4 In the class model 128, C0, Ci, C2, C3, and C4 are coefficients determined based on data of rolled materials belonging to the same class. As described above, in the invention, the class is a unit in which rolled materials are classified based on the similarity of the chemical composition, and typically the steel types are the same. However, the invention is not limited thereto, and a plurality of steel types may be classified into one class, or conversely, one steel type may be divided into a plurality of classes. Therefore, the coefficients C0, Ci, C2, C3, and C4 of the class model 128 of the deformation resistance in the invention are determined using performance data of rolled materials having similar chemical compositions. The coefficients C0, Ci, C2, C3, and C4 of the class model 128 are functions of alloy components of the rolled material, and can be expressed by a mathematical model shown in the following equation (14) in the same form as the coefficients of the global model 127 as one simple embodiment. C(3 = ^O.const + ^0,C Wc + C0 Si WSi + C0 Mn WMn +... ^1 = ^1,const + ^1,C Wc+ C1 Si WSi+ C1jMn WMn +... ^2 = ^2,const + ^2,C Wc + C2Sj WSi + C2Mn WMn +... ^3 = ^3,const + ^3,C Wc + C3Sj WSi + C3Mn WMn +... c4 = c4 const + c4 c wc + c4 Sj wSi + c4 Mn wMn +... (14 j Although only the terms of C, Si, and Mn are shown in the mathematical model of the coefficients of the class model 128, the invention is not limited thereto. In addition to the above terms, the mathematical model of the coefficient of the class model 128 may include terms of nickel (Ni), chromium (Cr), molybdenum (Mo), niobium (Nb), titanium (Ti), vanadium (V), nitrogen (N), aluminum (Al), copper (Cu), tungsten (W), and boron (B). The coefficients C0, Ci, C2, C3, and C4 of the class model 128 are functions of alloy components including a plurality of alloy component terms as shown in the above equation, and the contribution degree parameters C0,const, C0,c, and the like of each alloy component term are parameters determined using the rolling performance data 130 of rolled materials belonging to the corresponding class. A method of determining the parameters of the mathematical model of the coefficients of the class model 128 will be described later together with the description of an adjustment method. FIG. 4 shows an example of the rolling condition 110. For each number of the rolled material, a header portion 111 such as an alloy component or the like common to one rolled material, and a condition portion 112 including a temperature, a strain, and a strain rate different for each rolling pass and for each position of the rolled material are provided. The header portion 111 includes a material number 1101 for storing an identifier of a rolled material, a class 1102 for storing an identifier of a class set for the rolled material, a C 1103 for storing a content rate (or content) of carbon of the alloy component, a Si 1104 for storing a value of silicon of the alloy component, and a Mn 1105 for storing a value of manganese. The alloy component is not limited to the above, and a field can be set for each alloy component contained in the rolled material. The condition portion 112 includes, in one entry, a pass 1106, a position 1107, a temperature T 1108, a strain E 1109, and a strain rate sdot 1110. The position 1107 stores a position at which the state of the rolled material is measured in the hot rolling line 3. The position to be measured can be represented by a ratio of a length from a tip to the position in the total length of the rolled material. In the example of FIG. 4, the temperature (T 1108) , the strain (s 1109), and the strain rate (sdot 1110) at the positions of 5% (tip) , 50% (center) , and 95% (tail end) of the length of the rolled material in a longitudinal direction are input as the rolling condition for each pass 1106. Although FIG. 4 shows an example in which the rolling condition determination device 11 determines the class 1102 of the rolled material and inputs the determined class 1102 to the deformation resistance prediction device 12 as the rolling condition, the deformation resistance prediction device 12 may determine the class based on the alloy component of the rolled material. Although FIG. 4 shows an example in which the strain (E 1109) and the strain rate (sdot 1110) are calculated by the rolling condition determination device 11 and input to the deformation resistance prediction device 12 as the rolling condition, the rolling condition may include a rolling reduction amount, a rolling rate, a roll rotation speed, or the like for each pass instead of the strain and the strain rate, and the deformation resistance prediction device 12 may calculate the strain and the strain rate based om the rolling reduction amount (load), the rolling rate, or the like. Although not illustrated, the operation data 150 includes sensor data measured by a sensor provided in the hot rolling line 3, and can include, for example, a load, a temperature, a speed of the rolled material, a thickness of the rolled material, and a gap between rolls (rolled materials). The rolling performance data 130 is data obtained by combining the rolling condition 110 in FIG. 4 and the operation data 150, and includes, for each material number 1101 of the rolling condition 110, for example, the condition portion 112 of the rolling condition 110 and the sensor data of the operation data 150, which are not illustrated. Although the rolling performance data 130 can be accumulated for each rolled material for which rolling has been completed, the rolling performance data 130 may be statistical data obtained by performing statistical processing on a maximum value, an average value (median value) , a minimum value, or the like of sensor data in units of a predetermined number (or time) . The rolling performance data 130 may be generated based on the rolling condition 110 and the operation data 150 by the deformation resistance prediction device 12, or may be generated based on the rolling condition 110 and the operation data 150 by the rolling performance data storage device 13. FIG. 5 shows an example of the global model 127 stored in the deformation resistance prediction device 12 . The global model 127 includes a coefficient 1271, a Const 1272 for storing a constant, and fields (1273 to 1275) for storing a value corresponding to the alloy component. In the illustrated example, although examples of C, Si, and Mn are shown as alloy components, the alloy components are not limited thereto. The coefficient 1271 can store the identifier of the mathematical model as shown in the above equation (6) . The class model 128 stored in the deformation resistance prediction device 12 will be described later together with the description of an adjustment method. FIG. 6 shows the prediction accuracy evaluation step S2 of the invention. When the prediction accuracy evaluation step S2 shown in FIG. 2 is started, the deformation resistance prediction device 12 performs a rolling performance acquisition step S201 to acquire the rolling performance data 130 as an accuracy evaluation target from the rolling performance data storage device 13. Which rolling performance data 130 among the rolling performance data 130 is the accuracy evaluation target is determined by the setting of an execution frequency and an evaluation frequency of the prediction accuracy evaluation step S2 by the user. For example, when the execution frequency is set for each rolled material, regardless of the setting of the evaluation frequency, the deformation resistance prediction device 12 performs the prediction accuracy evaluation step S2 every time rolling is completed and the operation data 150 is stored for one rolled material, and acquires, as the accuracy evaluation target, the rolling performance data 130 of the rolled material rolled immediately before. On the other hand, when the execution frequency is set to every ten rolled materials and the evaluation frequency is set to each rolled material, the deformation resistance prediction device 12 performs the prediction accuracy evaluation step S2 every time the rolling of ten rolled materials is completed and the operation data 150 is stored, and acquires, as the accuracy evaluation target, the rolling performance data 130 of the ten rolled materials rolled immediately before. Alternatively, when the execution frequency is set to every ten rolled materials and the evaluation frequency is set to every two rolled materials, the deformation resistance prediction device 12 performs the prediction accuracy evaluation step S2 every time the rolling of ten rolled materials is completed and the operation data 150 is stored, and acquires, as the accuracy evaluation target, the rolling performance data 130 of two rolled materials of the ten rolled materials rolled immediately before. It is needless to say that various combinations of settings other than those exemplified above are also possible. Next, the deformation resistance prediction device 12 performs an abnormal data exclusion step S202 to determine the presence or absence of an abnormality in the temperature, the strain, the strain rate, and the like of the operation data 150, and excludes the data from the accuracy evaluation target when there is an abnormality. In order to determine the presence or absence of an abnormality, the deformation resistance prediction device 12 uses a difference between the operation data 150 and the rolling condition 110. For example, when a difference between the temperature of the operation data 150 and the temperature of the rolling condition 110 exceeds a predetermined threshold value and deviates from a normal range, the operation data 150 is determined to be abnormal . In this case, for example, a normal range may be set by the user, may be automatically calculated by applying statistical analysis to the rolling performance data 130, or may be determined by combining these two methods. Next, the deformation resistance prediction device 12 performs a global model application step S203, and calculates the global prediction value KPfG of the deformation resistance by applying the global model 127 of the deformation resistance stored in the deformation resistance prediction device 12 to the temperature, the strain, and the strain rate of the alloy component of the rolling condition 110 and the operation data 150. Next, the deformation resistance prediction device 12 performs a class model application step S204, calculates the class correction value APfC of the deformation resistance by applying the class model 128 of the deformation resistance stored in the deformation resistance prediction device 12 to the temperature, the strain, and the strain rate of the alloy component of the rolling condition 110 and the operation data 150, and calculates the prediction value Kp of the deformation resistance based on the correction value APfC and the global prediction value KPfG- As described above, the prediction value Kp can be calculated based on a sum or product of the correction value APfC and the global prediction value KPfG. Next, the deformation resistance prediction device 12 performs an accuracy evaluation step S205 to evaluate an accuracy of the deformation resistance prediction model 123. Therefore, the deformation resistance prediction device 12 compares the prediction value Kp of the deformation resistance calculated by applying the prediction model 123 with the actual value KPfact of the deformation resistance included in the operation data 150 (or the rolling performance data 130) to calculate the accuracy of the deformation resistance prediction model 123. As the actual value KPfact of the deformation resistance included in the operation data 150, a value calculated based on the operation data 150 by the rolling performance data storage device 13 can be used. Although an index of the accuracy is typically a 1/2 power of an average of squares of prediction values Kp-KPfact calculated at a plurality of positions in a plurality of passes for one rolled material, that is, root mean square error (RMSE), the index may be expressed by using a 1/2 power of an average of squares of (Kp-Kp,act) /KPfact, or other accuracy indexes. Alternatively, a plurality of accuracy indexes may be used. The deformation resistance prediction device 12 stores the prediction value Kp evaluated in the accuracy evaluation step S205, KPfact used for the above comparison, and one or more accuracy indexes in an accuracy history 124 in the deformation resistance prediction device 12 together with the number (material number 1101) , the class (1102) , and a rolling date and time of the rolled material. Next, the deformation resistance prediction device 12 performs an accuracy display step S206 to present the accuracy of the deformation resistance prediction model 123 to the user of the deformation resistance prediction device 12. In the accuracy display step S206, the accuracy index of the selected rolled material is read out from the deformation resistance prediction device 12 and presented to the user via an output device (described later). In the accuracy display step S206, the accuracy index to be displayed, the rolled material for displaying the accuracy index, and a display method of the accuracy index are changed according to a setting of the user, and the prediction accuracy of the deformation resistance prediction model 123 is displayed for each class, and then the processing ends. FIG. 7 shows an example of a screen 700 displayed on the output device in the accuracy display step S206. The screen 700 in FIG. 7 displays the RMSE of prediction values of the deformation resistance prediction model 123 of 100 rolled materials for each class evaluated most recently for six classes from a class 1001 to a class 1023. In the screen 700, a horizontal axis indicates a history of 100 evaluated rolled materials (-100) to the most recently most recently (0) , and a vertical axis indicates the RMSE of the deformation resistance prediction model 123 evaluated for each rolled material. FIG. 7 is an example of accuracy display, and other various display methods are also possible. For example, by setting the rolling date and time on the horizontal axis, a temporal change of the accuracy index can be displayed. Further, the accuracy index can be displayed more analytically by setting the vertical axis to the RMSE of the prediction value at a specific pass or the RMSE at a specific position instead of the RMSE of the prediction values at a plurality of passes and a plurality of positions in one rolled material. By displaying a moving average for a plurality of rolling materials instead of displaying the accuracy index for each rolling material, the display can be made smoother and a trend of change can be easily understood. FIG. 8 shows an example of a display obtained by making a moving average the RMSE shown in FIG. 7 with ten rolled materials. When FIG. 7 is viewed carefully, an increase in the RMSE of the class 1002 is also found, but it is easier to find an increase in the RMSE of the class 1002 in FIG. 8. The model adjustment necessity determination step S3 shown in FIG. 2 is performed when the above accuracy index is larger than a predetermined threshold value. The threshold value may be set by the user, or may be determined by statistically analyzing the data of the accuracy index of each class by the deformation resistance prediction device 12. FIG. 9 shows an example of a screen 900 on which the deformation resistance prediction device 12 determines a threshold value by statistical analysis and determines whether model adjustment is necessary. In the example of FIG. 9, the deformation resistance prediction device 12 defines the threshold value as Avg+2*std_dev using the moving average Avg and a moving standard deviation std_dev of the RMSE in the rolled material evaluated most recently. The RMSE for each rolled material shown by the solid line in the figure increases beyond a threshold value line (Avg+2*std_dev) from the horizontal axis = "-20", that is, from 20 rolled materials recently evaluated. In such a case, the deformation resistance prediction device 12 determines that the model adjustment is necessary, and issues a signal for prompting the user to perform the adjustment. The signal is, for example, screen display, sound, vibration, or electronic mail. When it is determined that the model adjustment is necessary, the deformation resistance prediction device 12 may execute the class model adjustment step S4 instead of or in addition to issuing a signal for prompting the user to perform the adjustment. FIG. 10 shows an example of a screen 950 on which model adjustment is determined to be unnecessary. In the screen 950, since the RMSE of each rolled material of the class 1001 is stopped in a region lower than the threshold value defined as Avg+2*std_dev, the deformation resistance prediction device 12 determines that the model adjustment is unnecessary. Although FIGS. 9 and 10 show examples in which Avg+2*std_dev is used as the threshold value for determining the necessity of model adjustment, other threshold values based on statistical analysis may be used. FIG. 11 shows the class model adjustment step S4 of the invention. When the class model adjustment step S4 is started, the deformation resistance prediction device 12 executes an adjustment class selection step S401 to select a class to be adjusted. The class to be adjusted may be selected by the deformation resistance prediction device 12 based on the accuracy index and the threshold value for determining the necessity of the model adjustment, may be selected by the user, or may be finally selected by the user by presenting a recommended class by the deformation resistance prediction device 12 to the output device. The deformation resistance prediction device 12 performs the steps of a class data extraction step S402 to a class model storage step S406 for each of the classes to be adjusted selected in the class data extraction step S402, and adjusts the class model 128 of the deformation resistance for each class. An adjustment end determination S407 regarding the selected class is processing for adjusting all the selected classes to be adjusted. Therefore, in the adjustment end determination S407, it is determined whether the processing for all the selected classes to be adjusted has been completed, and when there remains a class that has not been adjusted, the processing is returned to the class data extraction step S402 to continue the adjustment for the unadjusted class. Hereinafter, the above-described steps (S402 to S406) executed by the deformation resistance prediction device 12 for each class to be adjusted will be described. First, the deformation resistance prediction device 12 executes the class data extraction step S402 to acquire the rolling performance data 130 including the rolling condition 110 and the operation data 150 belonging to the class to be adjusted from the rolling performance data storage device 13 . Next, the deformation resistance prediction device 12 executes an abnormal data exclusion step S403 to exclude abnormal data from the acquired rolling performance data 130. Processing in this step may be the same as the abnormal data exclusion step S202 executed in the prediction accuracy evaluation step S2. Next, the deformation resistance prediction device 12 executes a global model application step S404, and calculates the global prediction value KPfG of the deformation resistance by 36 applying the global model 127 of the deformation resists in the deformation resistance prediction device temperature, strain, and strain rate of the alloy compc to 1105) of the rolling condition 110 and the operatioi Next, the deformation resistance prediction executes a class model adjustment step S405 to 6 parameters included in the mathematical model of the co of the class model 128. Hereinafter, as an example ol model 128, Ap c = exp( C0 + C-, In(Edot) + C2/T) x z03 (15) the processing in the class model adjustment ste] be described by using the above equation. The number of data of the rolling performance c the class used in the global model application step SA the abnormal data exclusion step S403 is assumed to be M. that the actual value of the deformation resistance data among the M pieces of data is Kp,act,m and the global value of the deformation resistance is Kp,G,m, it is desJ are calculated and the equation is organized using the example of the class model 128 of the deformation resistance, the following equation (17) is obtained. ln(Kpiactim/Kp,Gjn) =00 + 0! ln(£dotm) + C2/Tm + C3£m (17) Here, the coefficients C0, Ci, C2, and C3 of the class model 128 are functions of the alloy components as described above, and a simple example is a linear function of the alloy component as described in the following equation (18). Co = Co;Const + ^°>c ^c+ ^°.Si ^Si + ^0,Mn WMn +... ^1 = ^1,const + ^1,C Wc + C1jSj WSi + C1|Mn WMn +... ^2 = ^2,const + ^2,0 Wc + C2Si WSi + C2jMn WMn +... ^3 = ^3,const + ^3,C Wc + C3 Si WSi + C3 Mn WMn +... (1Q ) When organization is made using the linear function described above, a relationship between the mth performance data and the parameters included in the mathematical model of the coefficients of the class model 128 is expressed by the following equation (19) . ln(KPjact,m^KpG)m) = ^O.const + ^0,C WCm + C0Sj WSj m + ... + C0 E!m_N WE|m_N m + (^1,const+ ^i,c WCm + C1jSj WSjm + ... + C^Eim_N WEim_N>m) ln(£dotm) + (^2,const + ^2,C Wc>m + C2 SJ WSj>m + ... + C2 E|m_N WE|m_N m)/Tm + (C3,const + C3,C Wc,m + C3si Wsi,m + ... + C3, B^N WElm]N]m)£m (19) All of the M pieces of rolling performance data 130 used in the class model adjustment step S405 can be represented by an expression of the same form. When the M expressions are expressed in a form of a matrix, the following equation is obtained. Z=UXS (20) Here, Z on the left side of Equation (20) is a vector of M rows, and the mth row thereof is represented by ln(Kp,act,m/ Kp,G,m) <21> In the above equation (20) , U on the right side is a matrix of M rows and 4N + 4 columns, and the mth row thereof is given by the following equation (22). Um){1)...,N+1} = 1» WCm, WSim, ..., WE|m_Nm Um,{N+2,...,2N+2> = in(£dotm), ln(£dotm)Wc!m, !n(£dotm)WSim, ..., ln(£dotm) WElmNm Um,{2N+3,...,3N+3} = 1^"m» WCm/Tm, WSj>m/Tm, ..., WE|m_N>m/Tm Um>{3N+4,-,4N+4} = £m« Wc m£m, WSi m£m, ..., WE)m_N m£m (22) In the above equation (20) , S on the right side is a vector of 4N + 4 rows including parameters of a mathematical model of a class coefficient, and an element thereof is calculated by the following equation (23) . ${1,...,N+1} = C(),const> ^O.C ^0,Sh •••> ^0,Elm_N S{N+2,...,2N+2} = Ci)COnst» Cl,C> ^1,Si» • •> ^1,Elm_N ${2N+3,...,3N+3} = ^2,const. ^2,C> ^2,SM •■■> ^2,Elm_N (23) Therefore, the parameter of the mathematical model of the class coefficient is obtained by calculating the vector S at which the square error |Z-U> Although in the above embodiment, an example in which the deformation resistance prediction device 12 is applied to the hot rolling line control system 1 has been described, the invention is not limited thereto. The deformation resistance prediction device 12 can be applied to a warm rolling line or a cold rolling line instead of the hot rolling line 3 of the above embodiment. As described above, the hot rolling line control system 1 (or the deformation resistance prediction system) of the above embodiment may have the following configuration. (1) A deformation resistance prediction system, including: a rolling condition determination device (11) configured to determine a rolling condition (110) to be set in a rolling device (hot rolling line 3); a storage device (rolling performance data storage device 13) configured to collect operation data (150) of rolling performed by the rolling device (3) and the rolling condition (110); and a deformation resistance prediction device including a processor (21) and a memory (22) and configured to predict a deformation resistance of a rolled material based on the operation data (150) and the rolling condition (110) , in which the deformation resistance prediction device (12) includes a rolling data acquisition unit (12101) configured to acquire the operation data (150) and the rolling condition (110), accumulate the operation data (150) and the rolling condition (110) as rolling performance data (130), and acquire a rolling condition (110) candidate from the rolling condition (110) determination device (11), a prediction model (123) configured to predict the deformation resistance of the rolled material, and a deformation resistance prediction unit (121) configured to calculate a prediction value of the deformation resistance of the rolled material using the prediction model (123), the prediction model (123) includes a global model (127) determined based on the entire rolling performance data (130) and set in advance as a prediction model for estimating a deformation resistance of each steel type of the rolled material, and a class model (128) determined based on rolling performance data (130) of a deformation resistance of a specific steel type in which alloy components of the rolled material are similar or common among the rolling performance data (130) and set in advance as the prediction model for estimating the deformation resistance of the specific steel type, and the deformation resistance prediction unit (121) calculates a prediction value (Kp) of the deformation resistance of the rolled material based on the rolling condition (110) candidate using the global model (123) . With the above configuration, the deformation resistance is predicted by overlapping the global model of the deformation resistance determined by using the rolling performance data of a large number of steel types and the class model of the deformation resistance determined by using the rolling performance data of one or a small number (specific) of steel types, and thereby even for a steel type having a relatively small amount of rolling performance data, it is possible to predict the deformation resistance at least with an accuracy equal to or higher than an accuracy of the global model of the deformation resistance. (2) The deformation resistance prediction system according to (1) , in which the deformation resistance prediction unit (121) includes a global model application unit(12103) configured to apply the global model (127) to the rolling condition candidate (113) to calculate a global prediction value of the deformation resistance of the rolled material under the rolling condition candidate (113), and a class model application unit (12104) configured to apply the class model (128) to the rolling condition candidate (113), calculate a class correction value of the deformation resistance of the rolled material under the rolling condition candidate (113), and calculate a prediction value of the deformation resistance under the rolling condition (110) candidate based on the class correction value and the global prediction value. With the above configuration, the deformation resistance prediction device 12 can capture the general (global) tendency of the deformation resistance of the rolled material by the global model 127, capture the tendency of the deformation resistance in a class (specific steel type) different from the general tendency by the class model 128, and calculate the prediction value of the deformation resistance, and can provide, to the rolling condition determination device 11, the prediction value of the deformation resistance with a high accuracy. (3) The deformation resistance prediction system according to (1) , in which the deformation resistance prediction unit (121) further includes an accuracy evaluation unit (12106) configured to calculate a prediction accuracy of the prediction model based on the prediction value of the deformation resistance and an actual value of the deformation resistance included in the rolling performance data (130) . With the above configuration, it is possible to determine whether the class model needs to be adjusted by evaluating the prediction accuracy of the prediction model. (4) The deformation resistance prediction system according to (3), further including an accuracy display unit (12107) configured to generate a change history of a prediction accuracy of the class model (128) based on the prediction accuracy of the prediction model (123). With the above configuration, it is possible to determine whether the class model needs to be adjusted by grasping the change in the prediction accuracy of the prediction model in time series. (5) The deformation resistance prediction system according to (1) , in which the deformation resistance prediction unit (121) further includes a class model adjustment unit (12108) configured to adjust the class model (128) based on an actual value of the deformation resistance of the specific steel type included in the rolling performance data (130) . With the above configuration, the class model 128 can be adjusted based on the rolling result data 130 of the steel type corresponding to the class model 128. (6) The deformation resistance prediction system according to (1), in which the class model adjustment unit (12108) adjusts the class model (128) using the rolling performance data (130) of the specific steel type for the selected class model (128) . (7) The deformation resistance prediction system according to (3) , in which the deformation resistance prediction unit (121) further includes an abnormal data exclusion unit configured to exclude, as abnormal data, data in which a value of the rolling performance data (130) exceeds a predetermined range, and the accuracy evaluation unit (12106) calculates the prediction accuracy of the prediction model based on the rolling performance data (130) excluding the abnormal data. With the above configuration, the accuracy evaluation unit 12106 can accurately calculate the prediction accuracy by calculating the prediction accuracy of the prediction model based on the rolling performance data 130 excluding the abnormality data. (8) The deformation resistance prediction system according to (1), in which the global model (127) is configured by multiplying two factors, a first factor of the two factors being configured to include an addition result of a constant term, a logarithmic term of a strain rate of the rolling, and a power term of a temperature, and a second factor of the two factors being configured to include a power of a strain of the rolling. With the above configuration, the global model of the deformation resistance can be expressed based on the rolling performance data 130 of a large number of steel types. (9) The deformation resistance prediction system according to (1), in which the class model (128) is configured by multiplying two factors, a first factor of the two factors being configured by an exponential function having an exponential part including an addition result of a constant term, a logarithmic term of a strain rate of the rolling, and a power term of a temperature, and a second factor of the two factors being configured to include a power of a strain of the rolling. With the above configuration, the global model of the deformation resistance can be expressed based on the rolling performance data 130 of specific steel types. The invention is not limited to the above embodiments, and includes various modifications. For example, the above embodiments are described in detail for facilitating understanding of the invention, and are not necessarily limited to those including all the described configurations. A part of a configuration of one embodiment can be replaced with a configuration of another embodiment, and a configuration of another embodiment can be added to a configuration of one embodiment. With respect to a part of the configuration of each embodiment, addition, deletion, or replacement of another configuration can be applied alone or in combination. A part or all of the configurations, functions, processing units, processing methods or the like described above may be implemented by hardware such as through design using an integrated circuit. The above configurations, functions, or the like may be implemented by software by means of a processor interpreting and executing a program for implementing respective functions. Information of a program, a table, a file, etc. for realizing each function can be placed in a recording device such as a memory, a hard disk, or a solid state drive (SSD) , or in a recording medium such as an IC card, an SD card, or a DVD. Control lines or information lines indicate what is considered necessary for description, and not all the control lines or information lines are necessarily shown in a product. It may be considered that almost all the configurations are actually connected to each other. WE CLAIM: 1. A deformation resistance prediction system, comprising: a rolling condition determination device configured to determine a rolling condition to be set in a rolling device; a storage device configured to collect operation data of rolling performed by the rolling device and the rolling condition; and a deformation resistance prediction device including a processor and a memory and configured to predict a deformation resistance of a rolled material based on the operation data and the rolling condition, wherein the deformation resistance prediction device includes a rolling data acquisition unit configured to acquire the operation data and the rolling condition, accumulate the operation data and the rolling condition as rolling performance data, and acquire a rolling condition candidate from the rolling condition determination device, a prediction model configured to predict the deformation resistance of the rolled material, and a deformation resistance prediction unit configured to calculate a prediction value of the deformation resistance of the rolled material using the prediction model, the prediction model includes a global model determined based on the entire rolling performance data and set in advance as a prediction model for estimating a deformation resistance of each steel type of the rolled material, and a class model determined based on rolling performance data of a deformation resistance of a specific steel type in which alloy components of the rolled material are similar or common among the rolling performance data and set in advance as the prediction model for estimating the deformation resistance of the specific steel type, and the deformation resistance prediction unit calculates the prediction value of the deformation resistance of the rolled material based on the rolling condition candidate using the global model and the class model. 2. The deformation resistance prediction system according to claim 1, wherein the deformation resistance prediction unit includes a global model application unit configured to apply the global model to the rolling condition candidate to calculate a global prediction value of the deformation resistance of the rolled material under the rolling condition candidate, and a class model application unit configured to apply the class model to the rolling condition candidate, calculate a class correction value of the deformation resistance of the rolled material under the rolling condition candidate, and calculate a prediction value of the deformation resistance under the rolling condition candidate based on the class correction value and the global prediction value. 3. The deformation resistance prediction system according to claim 1, wherein the deformation resistance prediction unit further includes an accuracy evaluation unit configured to calculate a prediction accuracy of the prediction model based on the prediction value of the deformation resistance and an actual value of the deformation resistance included in the rolling performance data. 4. The deformation resistance prediction system according to claim 3, further comprising: an accuracy display unit configured to generate a change history of a prediction accuracy of the class model based on the prediction accuracy of the prediction model. 5. The deformation resistance prediction system according to claim 1, wherein the deformation resistance prediction unit further includes a class model adjustment unit configured to adjust the class model based on an actual value of the deformation resistance of the specific steel type included in the rolling performance data. 6. The deformation resistance prediction system according to claim 5, wherein the class model adjustment unit adjusts the class model using the rolling performance data of the specific steel type for the selected class model. 7. The deformation resistance prediction system according to claim 3, wherein the deformation resistance prediction unit further includes an abnormal data exclusion unit configured to exclude, as abnormal data, data in which a value of the rolling performance data exceeds a predetermined range, and the accuracy evaluation unit calculates the prediction accuracy of the prediction model based on the rolling performance data excluding the abnormal data. 8. The deformation resistance prediction system according to claim 1, wherein the global model is configured by multiplying two factors, a first factor of the two factors being configured to include an addition result of a constant term, a logarithmic term of a strain rate of the rolling, and a power term of a temperature, and a second factor of the two factors being configured to include a power of a strain of the rolling. 9. The deformation resistance prediction system according to claim 1, wherein the class model is configured by multiplying two factors, a first factor of the two factors being configured by an exponential function having an exponential part including an addition result of a constant term, a logarithmic term of a strain rate of the rolling, and a power term of a temperature, and a second factor of the two factors being configured to include a power of a strain of the rolling. 10. A deformation resistance prediction method in which a deformation resistance prediction device including a processor and a memory predicts a deformation resistance of a rolled material, the method comprising: a first step in which the deformation resistance prediction device acquires a rolling condition candidate from a rolling condition determination device configured to determine a rolling condition to be set in a rolling device; a second step in which the deformation resistance prediction device acquires, from a storage device configured to collect operation data of rolling performed by the rolling device and the rolling condition, the operation data and the rolling condition, and accumulates the operation data and the rolling condition as rolling performance data; and a third step in which the deformation resistance prediction device calculates a prediction value of the deformation resistance of the rolled material using a prediction model for predicting the deformation resistance of the rolled material, wherein in the third step, the prediction value of the deformation resistance of the rolled material is calculated based on the rolling condition candidate using the prediction model including a global model determined based on the entire rolling performance data and set in advance as a model for estimating a deformation resistance of each steel type of the rolled material, and a class model determined based on rolling performance data of a deformation resistance of a specific steel type in which alloy components of the rolled material are similar or common among the rolling performance data and set in advance as a model for estimating the deformation resistance of the specific steel type. 11. A deformation resistance prediction device that includes a processor and a memory and configured to predict a deformation resistance of a rolled material, the deformation resistance prediction device comprising: a rolling data acquisition unit configured to acquire operation data of rolling and a rolling condition, accumulate the operation data and the rolling condition as rolling performance data, and receive a rolling condition candidate; a prediction model configured to predict a deformation resistance of the rolled material; and a deformation resistance prediction unit configured to calculate a prediction value of the deformation resistance of the rolled material using the prediction model, wherein the prediction model includes a global model determined based on the entire rolling performance data and set in advance as a prediction model for estimating a deformation resistance of each steel type of the rolled material, and a class model determined based on rolling performance data of a deformation resistance of a specific steel type in which alloy components of the rolled material are similar or common among the rolling performance data and set in advance as the prediction model for estimating the deformation resistance of the specific steel type, and the deformation resistance prediction unit calculates the prediction value of the deformation resistance of the rolled material based on the rolling condition candidate using the global model and the class model.