Technical Field [0001] The embodiments herein generally relate to a method of predicting behavior of
a capital market, and more particularly, to a system and method for scalable predictive
analytics to determine a pattern that includes a capital market data over time for a capital
market, using a machine learning model executed on a server.
Description of the Related Art [0002] A capital market is a market where buyers and sellers engage in trade of
financial securities like stocks, commodities, bond etc. The buying/selling is undertaken by
participants such as individuals and institutions. Some of instruments which are traded in
capital market are index futures, index options, stocks, stocks futures, stocks options,
commodities, bonds etc. Nowadays, trading in capital market requires lot of technical
knowledge, due to this problem lot of people such as common man tend to lose money and
have negative impression about investing in capital market. Hence, there is a need for a
method to provide insights to enable a common man to walk through the trade market
without exposure towards complex trading strategies.
[0003] There are several methods available in the market to guide a user on trade market. In one approach, the method provides technical analysis to predict the future of the market by combining indicators of trend, momentum and volatility. However, this method provides only high probable indication that the next move of the market. These methods cannot predict the entire day’s movement. Other approaches may require understanding of technical parameters, charts, etc. and are complex to comprehend for quicker decisions.
[0004] Accordingly, there remains a need for a system and method for scalable
predictive analytics to determine a pattern that includes a capital market data over time for a
2

capital market, using a machine learning model executed on a server for entire day and for the subsequent days.
SUMMARY [0005] In view of a foregoing, an embodiment herein provides a method for scalable predictive analytics to determine a pattern that includes a capital market data over time for a capital market, using a machine learning model executed on a server. The method includes (i) generating a first database with a real time capital market data for every N minute associated with a first capital market, (ii) normalizing the real time capital market data associated with the first capital market by dividing real time capital market data for every N minute using a normalization factor (iii) providing the normalized real time capital market data associated with the first capital market into the machine learning model, wherein the machine learning model is generated by (a) generating a second database with a historical capital market data for every N minute associated with one or more capital markets, (b) normalizing the historical capital market data associated with one or more capital markets using a normalization factor, (c) processing a financial expert input on the normalized historical capital market data and (d) providing the historical capital market data associated with the one or more capital markets, the financial expert input on the normalized historical capital market data, to the machine learning model as training data to generate the machine learning model, (iv) recognizing, using the machine learning model, a first pattern that matches the selected portion of the normalized real time capital market data associated with the first capital market with the selected portion of the normalized historical capital market data associated with the one or more capital markets, and ranks the first pattern of the historical capital market data in an order of the match, (v) determining, a second pattern that includes a predicted first capital market data for a forthcoming period of the first capital market by combining the real time capital market data from a starting time to a present time

of the first capital market with the historical capital market data from the present time to one or more capital market closing time and (vi) determining a maximum rise percentage and a maximum fall percentage of the capital market from the predicted first capital market data for the forthcoming period of the first capital market from the second pattern and obtaining a risk ratio by dividing the maximum rise percentage and the maximum fall percentage for recommending an action that includes at least one of BUY or SELL a capital market asset associated with the first capital market based on the obtained risk ratio. The normalization factor for normalizing the real time capital market data is obtained by dividing a first starting point of the real time capital market data of a day by a constant number. The constant is any high positive number that is nearest to the real capital market data at the starting of the day. The selected portion includes a capital market data from the starting time to a present time of the capital market.
[0006] In one embodiment, the method further comprises recommending to BUY the capital market asset associated with the first capital market when the risk ratio is greater than 3.
[0007] In another embodiment, the method further comprises recommending to SELL the capital market asset associated with the first capital market when the risk ratio is less than 3.
[0008] In yet another embodiment, the real time capital market data and the historical capital market data includes at least one of an open price, a high price, a low price or a close price of the capital market asset.
[0009] In yet another embodiment, the historical capital market data associated with the one or more capital markets is normalized by dividing the historical capital market data for every N minute using a normalization factor. The normalization factor for normalizing historical capital market data is obtained by dividing a first starting point of the capital

market of the day by a constant number and the constant number is any high positive number that is nearest to the one or more capital market data at the starting of the day.
[0010] In yet another embodiment, the method further comprises enabling a user to SELL or BUY the capital market asset associated with the first capital market by providing a user interface.
[0011] In yet another embodiment, the method further comprises automatically correcting the recommendation when the recommendation is incorrect by recognizing a third pattern that matches selected portion of the normalized real time capital market data associated with the first capital market with the selected portion of the normalized historical capital market data associated with the one or more capital markets and predicting modified first capital market data for the forthcoming period of the first capital market by combining the real time capital market data from a starting time to a present time of the first capital market with the historical capital market data from the present time to one or more capital market closing time. The selected portion includes the capital market data from the starting time to the present time of the market.
[0012] In yet another embodiment, the method further comprises performing the scalable predictive analytics for the first capital market for subsequent days.
[0013] In yet another embodiment, the capital market asset includes at least one of stock or index.
[0014] In one aspect, a system including a server for performing scalable predictive analytics to determine a pattern that includes capital market data over time for a capital market, using a machine learning model is provided. The system includes a memory that stores program codes and a processor is configured to: (i) generate a first database with a real time capital market data for every N minute associated with a first capital market, (ii) normalize the real time capital market data associated with the first capital market by dividing

the real time capital market data for every N minute using a normalization factor, (iii) provide the normalized real time capital market data associated with the first capital market into the machine learning model, wherein the machine learning model is generated by (a) generating a second database with a historical capital market data for every N minute associated with one or more capital markets, (b) normalizing the historical capital market data associated with the one or more capital markets, (c) processing a financial expert input on the normalized historical capital market data; and (d) providing (a) the historical capital market data associated with the one or more capital market, (b) the financial expert input on the normalized historical capital market data, to the machine learning model as training data to generate the machine learning model, (iv) recognize, using the machine learning model, a first pattern that matches the selected portion of the normalized real time capital market data associated with the first capital market with the selected portion of the normalized historical capital market data associated with the one or more capital markets and rank the first pattern of the historical capital market data in an order of the match. (v) determine, a second pattern that includes a predicted first capital market data for a forthcoming period of the first capital market by combining the normalized real time capital market data from a starting time to a present time of the first capital market with the normalized historical capital market data from the present time to one or more capital market closing time, (vi) determine a maximum rise percentage and a maximum fall percentage of the capital market from the predicted first capital market data for the forthcoming period of the first capital market from the second pattern and obtaining a risk ratio by dividing the maximum rise percentage and the maximum fall percentage for recommending an action that includes at least one of BUY or SELL a capital market asset associated with the first capital market based on the obtained risk ratio. The normalization factor for normalizing the real time capital market data is obtained by dividing a first starting point of the real time capital market data of a day by a constant

number. The constant is any high positive number that is nearest to the real capital market data at the starting of the day. The selected portion includes a capital market data from the starting time to a present time of the capital market.
[0015] In one embodiment, the processor is further configured to recommend to BUY the capital market asset associated with the first capital market when the risk ratio is greater than 3.
[0016] In another embodiment, the processor is further configured to recommend to SELL the capital market asset associated with the first capital market when the risk ratio is less than 3.
[0017] These and other aspects of the embodiments herein will be better appreciated and understood when considered in conjunction with the following description and the accompanying drawings. It should be understood, however, that the following descriptions, while indicating preferred embodiments and numerous specific details thereof, are given by way of illustration and not of limitation. Many changes and modifications may be made within the scope of the embodiments herein without departing from the spirit thereof, and the embodiments herein include all such modifications.
BRIEF DESCRIPTION OF THE DRAWINGS
[0018] The embodiments herein will be better understood from the following detailed description with reference to the drawings, in which:
[0019] FIG. 1 is a system view of a system that predicts a pattern of capital market according to an embodiment herein;
[0020] FIG. 2 is a flow chart illustrating a method of predicting the pattern of capital market according to one embodiment herein;
[0021] FIGS. 3A and 3B are flow chart illustrating a method of predicting a pattern of capital market according to another embodiment herein;

[0022] FIG. 4 is a flow chart illustrating a method of generating machine learning algorithm, according to an embodiment herein; and
[0023] FIG. 5A and 5B are a graphical representation of a predicted pattern of capital market according to one embodiment herein.
DETAILED DESCRIPTION OF PREFERRED EMBODIMENTS
[0024] The embodiments herein and the various features and advantageous details thereof are explained more fully with reference to the non-limiting embodiments that are illustrated in the accompanying drawings and detailed in the following description. Descriptions of well-known components and processing techniques are omitted so as to not unnecessarily obscure the embodiments herein. The examples used herein are intended merely to facilitate an understanding of ways in which the embodiments herein may be practiced and to further enable those of skill in the art to practice the embodiments herein. Accordingly, the examples should not be construed as limiting the scope of the embodiments herein.
[0025] There remains a need for a system and method for scalable predictive analytics to determine a pattern that includes a capital market data over time for a capital market, using a machine learning model executed on a server. Referring now to the drawings, and more particularly to FIGS. 1 through 5B, where similar reference characters denote corresponding features consistently throughout the figures, there are shown preferred embodiments.
[0026] FIG. 1 is a system view of a system that performs scalable predictive analysis to determine a pattern of a capital market according to an embodiment herein. In one embodiment, the pattern of the capital market includes a plot of capital market data over a time. In one embodiment, the capital market data includes at least one of an open price, a high price, a low price or a close price of a capital market asset. In another embodiment, the capital market asset includes at least one of stock or index. The system includes a real time

tick data capturing module 102, a memory data store 104, a historical tick data capturing module 106, a preprocessing module 108, a back testing module 110, an analysis engine 112, a job scheduler 114, an API service module 116 and a user interface 118. The real time tick data capturing module 102 captures real time tick data from a data service provider. The historical tick data capturing module 106 captures a historical tick data from a data service provider. The preprocessing module 108 sanity checks and normalizes the historical tick data and real time tick data. The back testing module 110 back tests a machine learning model on the historical tick data. The memory data store 104 stores processed real time tick data and historical tick data. The analysis engine 112 executes the machine learning model. The analysis engine 112 includes a pattern analysis module, a rules and ranking module, a prediction module, and a recommendation module that predicts a pattern of the capital market and providing recommendation to BUY or SELL the capital market asset based on the predicted a pattern of the capital market. The job scheduler 114 schedules job to find out matching pattern for real time tick data from the historical tick data for periodically. In one embodiment, the API service module 116 may be a web service. The user interface 118 displays charts and recommendation on the capital market which can be easily followed by user and enabling the user to query the system to get the pattern of capital market at a given time for a given date. In one embodiment, the real tick data is real time capital market data. In one embodiment, the historical tick data is historical capital market data. In one embodiment, the job scheduler 114 schedules a job to find out matching pattern for real time tick data from the historical tick data for every one minute. In one embodiment, the memory data store 104 includes a first database for storing processed real time tick data and a second database for storing processed historical tick data.
[0027] The pattern analysis module determines a first pattern that matches the selected portion of the normalized real time tick data with the selected portion of the

normalized historical tick data. The rules and ranking module ranks the first pattern of normalized historical tick data based on the order of the match. The prediction module generates a second pattern that includes a predicted first capital market data for a forthcoming period of the first capital market by combining the normalized real time tick data from a starting time to a present time of the first capital market with the normalized historical tick data from the present time to closing time of the one or more capital markets. The recommendation module determines a maximum rise percentage and a maximum fall percentage of the capital market asset from the predicted first capital market data for the forthcoming period of the first capital market from the second pattern and obtaining a risk ratio by dividing the maximum rise percentage and the maximum fall percentage for recommending an action that includes at least one of BUY or SELL the capital market asset associated with the first capital market based on the obtained risk ratio through the user interface 118.
[0028] FIG. 2 is a flow chart illustrating the method of prediction of the pattern for a capital market according to one embodiment herein. At step 202, the historical capital market data is captured from a data service provider. At step 204, the historical capital market data is normalized using a normalization factor. At step 206, a machine learning model is generated. At step 208, the machine learning model is trained using the normalized historical capital market data. At step 210, the real time capital market data is captured from a data service provider. At step 212, the real time capital market data is normalized on the fly using the normalization factor. At step 214, the machine learning model is finalized based on a best pattern match of the historical capital market data. At step 216, the machine learning model predicts a pattern of the capital market in real time. In one embodiment, the historical capital market data and the real time capital market data include an open price, a high price, a low price or a close price of a capital market asset

[0029] FIGS. 3A and 3B are flow diagrams illustrating a method for scalable predictive analytics to determine a pattern for a capital market, using a machine learning model executed on a server according to another embodiment herein. At step 302, a first database with a real time capital market data for every N minute associated with the first capital market is generated. At step 304, the real time capital market data is normalized using a normalization factor. At step, 306, the normalized real time capital market data is provided to the machine learning model. At step 308, the first pattern that matches the selected portion of the normalized real time capital market data associated with the first capital market for a day with the selected portion of the normalized historical capital market data associated with the one or more capital markets is recognized using a machine learning model. At step 310, the first pattern of the normalized historical capital market data associated with one or more capital markets is ranked based on order of the match. At step 312, a second pattern that includes a predicted first capital market data for a forthcoming period of the first capital market is determined by combining the normalized real time capital market data from a starting time to a present time of the first capital market with the normalized historical capital market data from the present time to closing time of the one or more capital markets. At step 314, a maximum rise percentage and a maximum fall percentage of the capital market asset is determined from the predicted first capital market data for the forthcoming period of the first capital market from the second pattern and a risk ratio is obtained by dividing the maximum rise percentage and the maximum fall percentage for recommending an action that includes at least one of BUY or SELL the capital market asset associated with the first capital market based on the obtained risk ratio. In one embodiment, the real time capital market data and historical capital market data are obtained from a first system and a second system respectively and the first system and the second system include a data provider. In one embodiment, the normalization factor is obtained by dividing a first starting point of the real

time capital market data of a day by a constant number. the constant number is any high positive number that is nearest to the real capital market data at the starting of the day. The selected portion includes a capital market data from the starting time to the present time of the capital market. The first pattern of the historical capital market data is ranked based on an order of the match. In one embodiment, the real time capital market data and the historical capital market data includes at least one of an open price, a high price, a low price or a close price of the first capital asset.
[0030] FIG. 4 illustrates a method of generating a machine learning model according to an embodiment herein. At step 402, a second database with a historical capital market data associated with one or more capital markets is generated. At step 404, the historical capital market data is normalized using a normalization factor. At step 406, a financial expert input that is obtained from an expert device is processed on the normalized historical capital market data. At step 408, (a) the historical capital market data associated with the one or more capital markets, (b) the financial expert input on the normalized historical capital market data are provided to the machine learning model as a training data. In one embodiment, the normalization factor is obtained by dividing a first starting point of capital market data of a day by a constant number. The constant number is any high positive number that is nearest to the real capital market data at the starting of the day. The constant number did not change even if the instrument value changes in future.
[0031] The method automatically corrects the recommendation when the recommendation is incorrect by recognizing a third pattern that matches a selected portion of the normalized real time capital market data associated with the first capital market with the selected portion of the normalized historical capital market data associated with the one or more capital markets and predicting modified first capital market data for the forthcoming period of the first capital market by combining the real time capital market data from a

starting time to a present time of the first capital market with the historical capital market data from the present time to one or more capital market closing time. The selected portion includes the capital market data from the starting time to the present time of the market.
[0032] The method further predicts a pattern of capital market for subsequent days.
[0033] FIG. 5A is a graphical representation of the predicted pattern of capital market according to an embodiment herein. In one embodiment, at about time 9:20 am the system predicts that the capital market data will go down hence a down arrow indicating at time 9.20 am to BUY PUT (the put option increase in value if capital market goes down). In one embodiment, at time 9.30 am, the system predicts that the capital market data will go up hence the position closed and is indicated by star mark on the graph. In one embodiment, at time 9.50 am, the system predicts that stock will go up but the prediction is incorrect and then at time 10:10 am the position is closed. In one embodiment, at time 10:10 am, the machine learning model auto corrects and predicts that the capital market data will go down and recommended BUY PUT. In one embodiment, at about time 11:20 am after making some profit, the system recommends to close the position as indicated in star mark. In one embodiment, at time 11:25 am, again the system predicts that market will go down and the tool recommends “BUY 10300 PUT” as indicated in box in the graph. Today’s gain section in the graph provides gain value todays trade.
[0034] With reference to the FIG. 5A, FIG. 5B is a graphical representation of predicted pattern of a capital market according to one embodiment herein. In one embodiment, at time 11.53 am, the system provides recommendation to close the position based on the prediction at time 11.25 am (as in FIG. 5A). Hence, the system asks the trader to close the position make at around time 11:25 am, by recommending square off put. Then the todays gain value is increased.
[0035] The method of predicting a pattern of a capital market according to an

embodiment herein is explained in the following example. The example used herein is intended merely to facilitate an understanding of ways in which the embodiments herein may be practiced and to further enable those of skill in the art to practice the embodiments herein. Accordingly, the example should not be construed as limiting the scope of the embodiments herein.
[0036] Example 01
[0037] Consider today is 28 May, 2018
[0038] Consider, the market opens at time 9.15 am
[0039] Consider, the present time is time 11.00 am
[0040] The machine learning model runs at time 11.00 am for predicting a behavior of a capital market for a day (28 May, 2018). The model does the following steps:
[0041] Step: 1 Capital market data from time 9.15 am to 11.00 am is considered as input data for the machine learning model
[0042] Step 2: The machine learning model starts to find the closest matching pattern for 28 May, 2018 in the historical capital market data.
[0043] Step 3: The machine learning model finds best matches for 28 May, 2018. Let’s consider, its 10 Feb 2017.
[0044] Step 4: From time 11.00 am to 3.30 pm (market close time) of Feb, 2017 is mapped to 28 May, 2018 at time 11.00 am.
[0045] Step 5: Now, the recommendation will be done based on the predicted graph from time 11.00 am to 3.30 pm.
[0046] Step 6: Step 1 to 5 is repeated for every one minute
[0047] The foregoing description of the specific embodiments will so fully reveal the general nature of the embodiments herein that others can, by applying current knowledge, readily modify and/or adapt for various applications such specific embodiments without

departing from the generic concept, and, therefore, such adaptations and modifications should and are intended to be comprehended within the meaning and range of equivalents of the disclosed embodiments. It is to be understood that the phraseology or terminology employed herein is for the purpose of description and not of limitation. Therefore, while the embodiments herein have been described in terms of preferred embodiments, those skilled in the art will recognize that the embodiments herein can be practiced with modification within the spirit and scope of the appended claims.

CLAIMS I/We Claim:
1. A method for scalable predictive analytics to determine a pattern that includes a capital market data over time for a capital market, using a machine learning model executed on a server, characterized in that the method comprises:
generating a first database with a real time capital market data for every N minute associated with a first capital market (302);
normalizing the real time capital market data associated with the first capital market by dividing the real time capital market data for every N minute using a normalization factor (304), wherein the normalization factor is obtained by dividing a first starting point of the real time capital market data of a day by a constant number, wherein the constant number is any high positive number that is nearest to a first capital market data at the starting of the day;
providing the normalized real time capital market data associated with the first capital market into the machine learning model (306), wherein the machine learning model is generated by
generating a second database with a historical capital market data for every N minute associated with one or more capital markets (402);
normalizing the historical capital market data associated with the one or more capital markets (404);
processing a financial expert input on the normalized historical capital market data (406); and
providing (a) the historical capital market data associated with the one or more capital market, (b) the financial expert input on the normalized historical capital market data, to the machine learning model as training data to generate the machine learning model (408);

recognizing, using the machine learning model, a first pattern that matches the selected portion of the normalized real time capital market data associated with the first capital market for the day with the selected portion of the normalized historical capital market data associated with the one or more capital markets (308), wherein the selected portion comprises a capital market data from a starting time to a present time of the capital market, and ranks the first pattern of the historical capital market data in an order of the match (310);
determining, a second pattern that includes a predicted first capital market data for a forthcoming period of the first capital market by combining the normalized real time capital market data from a starting time to a present time of the first capital market with the normalized historical capital market data from the present time to one or more capital market closing time (312); and
determining a maximum rise percentage and a maximum fall percentage of the capital market from the predicted first capital market data for the forthcoming period of the first capital market from the second pattern and obtaining a risk ratio by dividing the maximum rise percentage and the maximum fall percentage for recommending an action that comprises at least one of BUY or SELL a capital market asset associated with the first capital market based on the obtained risk ratio (314).
2. The method as claimed in claim 1, comprises recommending to BUY the capital market asset associated with the first capital market when the risk ratio is greater than 3.
3. The method as claimed in claim 1, comprises recommending to SELL the capital market asset associated with the first capital market when the risk ratio is less than 3.

4. The method as claimed in claim 1, wherein the real time capital market data comprises at least one of an open price, a high price, a low price or a close price of the capital market asset.
5. The method as claimed in claim 1, wherein the historical capital market data comprises at least one of an open price, a high price, a low price or a close price of the one or more capital market asset.
6. The method as claimed in claim 1, wherein the historical capital market data associated with the one or more capital markets is normalized by dividing the historical capital market data for every N minute using the normalization factor, wherein the normalization factor is obtained by dividing a first starting point of historical capital market data of the day by a constant number, wherein the constant number is any high positive number that is nearest to the one or more capital market data at the starting of the day.
7. The method as claimed in claim 1, comprises enabling a user to SELL or BUY the capital market asset associated with the first capital market by providing a user interface.
8. The method as claimed in claim 1, comprises automatically correcting the recommendation when the recommendation is incorrect by recognizing a third pattern that matches selected portion of the normalized real time capital market data associated with the first capital market for the day with the selected portion of normalized historical capital market data associated with the one or more capital markets, wherein the selected portion includes the capital market data from the opening time to the present time of the capital market and predicting a modified first capital market data for the forthcoming period of the

first capital market by combining the real time capital market data from a first capital market starting time to a present time with the historical capital market data from the present time to one or more capital market closing time.
9. The method as claimed in claim 1, comprises performing the scalable predictive analytics for the first capital market for subsequent days.
10. The method as claimed in claim 1, wherein the capital market asset includes at least one of stock or index.
11. A system comprising a server for performing scalable predictive analytics to determine a pattern that includes capital market data over time for a capital market, using a machine learning model, the system comprising:
a memory that stores program codes; and a processor is configured to:
generate a first database with a real time capital market data for every N minute associated with a first capital market;
normalize the real time capital market data associated with the first capital market by dividing the real time capital market data for every N minute using a normalization factor, wherein the normalization factor is obtained by dividing a first starting point of the real time capital market data of a day by a constant number, wherein the constant number is any high positive number that is nearest to the real capital market data at the starting of the day;
provide the normalized real time capital market data associated with the first capital market into the machine learning model, wherein the machine learning model is generated by

generating a second database with a historical capital market data for every N minute associated with one or more capital markets;
normalizing the historical capital market data associated with the one or more capital markets;
processing a financial expert input on the normalized historical capital market data; and
providing (a) the historical capital market data associated with the one or more capital market, (b) the financial expert input on the normalized historical capital market data, to the machine learning model as training data to generate the machine learning model;
recognize, using the machine learning model, a first pattern that matches the selected portion of the normalized real time capital market data associated with the first capital market for the day with the selected portion of the normalized historical capital market data associated with the one or more capital markets, wherein the selected portion comprises a capital market data from a starting time to a present time of the capital market, and ranks the first pattern of the historical capital market data in an order of the match;
determine, a second pattern that includes a predicted first capital market data for a forthcoming period of the first capital market by combining the normalized real time capital market data from a starting time to a present time of the first capital market with the normalized historical capital market data from the present time to one or more capital market closing time; and
determine a maximum rise percentage and a maximum fall percentage of the capital market from the predicted first capital market data for the forthcoming period of the first capital market from the second pattern and obtaining a risk ratio by dividing the maximum rise percentage and the maximum fall percentage for recommending an action

that comprises at least one of BUY or SELL a capital market asset associated with the first capital market based on the obtained risk ratio.
12. The system as claimed in claim 11, wherein the processor is further configured to recommend to BUY the capital market asset associated with the first capital market when the risk ratio is greater than 3.
13. The system as claimed in claim 11, wherein the processor is further configured to recommend to SELL the capital market asset associated with the first capital market when the risk ratio is less than 3.