Description:Complete Specification:
The following specification describes and ascertains the nature of this invention and the manner in which it is to be performed

Field of the invention
[0001] The present disclosure relates to the field of data analytics and machine learning. In particular, the present invention discloses a system to predict impact on customer preferred optimization parameter of a product based on customer customizations and a training method thereof.

Background of the invention
[0002] Customers in the modern era enjoy choice in their purchase of a product and customized goods invite them to engage with a sales representative even more. With the rise in online retail and mass customization in manufacturing, businesses across every industry offer elements of customization for their products. Industries like automobile offer customers to configure the configuration of the product model from scratch. However, the customers are skeptical about making lot of decisions without knowing how it will be reflected in the end configuration from a product pricing perspective or any other customer preferred optimization parameter.

[0003] Introducing Intelligence in this process will aid the customer in configuring the product and enhance the overall customer experience by providing deep insights into the selection that they make. The intelligence that will be introduced will predict the impact of the customer’s selection of a particular configurable part on the preferred optimization parameter, such as but not limited to price, time taken for production, quality of the product. With the advent of data science and data analytics, it is convenient to use statistics and Machine learning concepts to make such predictions.

[0004] Research Paper titled “Information modeling and redundant data processing of mechanical product configuration design” IEEE, 10.1109/IMCEC51613.2021.9482305, 19 July 2021 studies the possibility of reducing the information that is shown in mechanical product configuration design information through redundant data processing. According to the characteristics of mechanical product configuration design information in the big data environment, the paper studies the construction of product configuration design information model and redundant data processing. Based on the characteristics of mechanical product configuration, design information and big data, it proposes a galaxy model suitable for mechanical product configuration, and design information. The paper further integrates this information into the galaxy model Structured data, unstructured data and semi-structured data that are expressed as JSON (JavaScript object notation) metadata. Using the sensitivity and semantic similarity of keyword data source a text similarity calculation method of JSON data file is proposed. Based on the text similarity calculation redundant data processing of mechanical product configuration design information is realized.

Brief description of the accompanying drawings
[0005] An embodiment of the invention is described with reference to the following accompanying drawings:
[0006] Figure 1 depicts a system (100) to predict impact on the customer preferred optimization parameters of a product based on user customizations;
[0007] Figure 2 illustrates method steps (200) of training an AI model (11) to predict impact on the customer preferred optimization parameters of a product;
[0008] Figure 3 depicts the user interface (20) displaying customer preferred optimization parameter for the product.

Detailed description of the drawings
[0009] Figure 1 depicts a system (100) to predict impact on customer preferred optimization parameter of a product based on customer customizations. The product could be any industrially manufactured product that is made of different parts/components. The customer preferred optimization parameter refers to production time, price, quality of the product etc. The customer customization refers to all possible combinations of different variants of the components that make up a complete product build (also called as bill of materials that make up the product). This disclosure is explained with a non-limiting example of a product from the automotive industry, the chassis of a vehicle. The various components of the chassis such as wheel configuration, wheel diameter, rim material, axle distance, suspension system and their variants are chosen by the customer. The system (100) comprises at least a user interface (20) in communication with a processor (10).

[0010] User interface (20) as used in this invention refers to a device that facilitates the point of human-computer interaction and communication in the system (100). This can include display screens in the form of a browser, keyboards, and the like. The user interface (20) can either be a single component comprising both the input and output interface or have two separate components the input interface and the output interface respectively. For example, the user interface (20) device could be any electronic or a cloud computing device from the group of a mobile phone unit, a computer, a tablet and the like. The user interface (20) is configured to receive customer configurations for the product and display the predicted impact of the customer preferred optimization parameter of the product.

[0011] The processor (10) can either be a logic circuitry or a software programs that respond to and processes logical instructions to get a meaningful result. A hardware processor (10) may be implemented as any or a combination of: one or more microchips or integrated circuits interconnected using a parent board, hardwired logic, software stored by a memory device and executed by a microprocessor, firmware, an application specific integrated circuit (ASIC), and/or a field programmable gate array (FPGA). In an exemplary embodiment of the present invention, the processor (10) runs a trained AI model (11). The AI model (11) is trained using method steps 200. In an alternate embodiment the AI model (11) is a separate component embodied in a separate module and in communication with the processor (10).

[0012] An AI model (11) can be defined as reference or an inference set of data, which uses different forms of correlation matrices. Using these models and the data from these models, correlations can be established between different types of data to arrive at some logical understanding of the data. A person skilled in the art would be aware of the different types of AI model (11)s such as linear regression, naïve bayes classifier, support vector machine, neural networks and the like. It must be understood that although this AI model (11) forms the core of this invention, this disclosure is not specific to the type of model being executed in the AI module and can be applied to any AI module irrespective of the AI model (11) being executed. A person skilled in the art will also appreciate that the AI model (11) may be implemented as a set of software instructions, combination of software and hardware or any combination of the same. For example, a neural network can be a software residing in the system (100) or the cloud or embodied within an electronic chip. Such neural network chips are specialized silicon chips, which incorporate AI technology and are used for machine learning.

[0013] The processor (10) is configured to retrieve a dataset comprising all possible combinations for a plurality of components of the product and the associated impact of the customer preferred optimization parameter for the product for said combination. The dataset can be stored in an internal or external memory. The processor (10) then applies feature engineering techniques on the dataset to determine a magnitude of impact each component has on each impact of the customer preferred optimization parameter of the product. The feature engineering techniques use either one or more from the group of T-test, Z-test, Pearson coefficient , Cohen values. The AI model (11) is configured to learn coefficients of each component for predicting the impact on customer preferred optimization parameter using the dataset, the determined magnitude of impact and the values of each component of the plurality of components. The processor (10) configured to calculate the customer preferred optimization parameter of user customized product by executing the AI model (11). The AI model (11) runs linear regression algorithms to predict impact of the customer preferred optimization parameter of the product.

[0014] As used in this application, the terms "component," "system (100)," "module," "interface," are intended to refer to a computer-related entity or an entity related to, or that is part of, an operational apparatus with one or more specific functionalities, wherein such entities can be either hardware, a combination of hardware and software, software, or software in execution. As further yet another example, interface(s) can include input/output (I/O) components as well as associated processing unit, application, or Application Programming Interface (API) components. A module with reference to this disclosure refers to a logic circuitry or a set of software programs that respond to and processes logical instructions to get a meaningful result. The system (100) could be a hardware combination of these modules or could be deployed remotely on a cloud or server.

[0015] It should be understood at the outset that, although exemplary embodiments are illustrated in the figures and described below, the present disclosure should in no way be limited to the exemplary implementations and techniques illustrated in the drawings and described below.

[0016] Figure 2 illustrates method steps (200) of training an AI model (11) to predict impact on customer preferred optimization parameter of a product based on customer customizations. The AI model (11) in communication with a processor (10) of a system (100). The system (100) further comprising a user interface (20) in communication with the processor (10). The system (100) and its constituents have been elaborated above in accordance with figure 1. It is reiterated that in an exemplary embodiment of the present invention, the processor (10) runs the AI model (11). In alternate embodiment the AI model (11) is a separate component embodied in a separate module and in communication with the processor (10).

[0017] Method step 201 comprises retrieving a dataset comprising all possible combinations for a plurality of components of the product and the associated impact of the customer preferred optimization parameter for the product for said combination. Table 1 below is non-exhaustive example of such dataset for the product, chassis of an automobile.

Wheel Configuration Wheel Diameter [in] Rim material Axle Distance Suspension system (100) Price ( Customer preferred optimization parameter)
6x4-4 20 Stainless steel 180 Leaf (Front +rear) 440
6x4-4 20 Stainless steel 180 Leaf (Front) + Air (rear) 450
6x4-4 20 Aluminum 180 Air (Front+rear) 460
6x4-4 22 Stainless steel 180 Leaf (Front +rear) 450
6x4-4 22 Aluminum 180 Leaf (Front) + Air (rear) 460
6x4-4 22.5 Stainless steel 180 Air (Front+rear) 480
6x4-4 22.5 Aluminum 180 Leaf (Front +rear) 460
6x4-4 24 Stainless steel 180 Air (Front+rear) 500
6x4-4 24 Aluminum 180 Leaf (Front +rear) 480
6x4 20 Stainless steel 180 Air (Front+rear) 440
6x4 20 Aluminum 180 Leaf (Front +rear) 420
6x2-4 22 Stainless steel 180 Air (Front+rear) 470
6x2-4 22 Aluminum 180 Leaf (Front +rear) 450
6x2-4 24 Stainless steel 180 Leaf (Front) + Air (rear) 490
6x2/4 20 Aluminum 180 Leaf (Front +rear) 410
6x2-2 22.5 Aluminum 180 Leaf (Front +rear) 420
6x2-2 22.5 Aluminum 180 Leaf (Front) + Air (rear) 430
6x2-2 22.5 Aluminum 180 Air (Front+rear) 440
6x2-2 24 Stainless steel 180 Air (Front+rear) 460
6x2-2 24 Aluminum 180 Leaf (Front +rear) 440
6x2 20 Aluminum 180 Air (Front+rear) 410
4x2 22.5 Stainless steel 180 Leaf (Front +rear) 400
4x2 22.5 Stainless steel 180 Leaf (Front) + Air (rear) 410

Table 1.
[0018] Method step 202 comprises applying feature engineering techniques on the dataset by means of the processor (10) to determine a magnitude of impact each component has on the customer preferred optimization parameter of the product. The feature engineering techniques use either one or more from the group of T-test, Z-test, Pearson coefficient , Cohen values. The Z-Test and T-test tell us whether or not a component has impact on the customer preferred optimization parameter or not. Then Pearson coefficient and Cohen values determined to give us the magnitude of such influence.

[0019] Method step 203 comprises coefficients of each component for predicting the impact on customer preferred optimization parameter using the dataset, the determined magnitude of impact and the values of each component of the plurality of components by means of the AI model (11). The impact of the customer preferred optimization parameter coefficients are different for each components w.r.t to each impact of the customer preferred optimization parameter. Further the values of these customer preferred optimization parameter coefficients are dynamic and changes as the AI model (11) learns on the plurality of datasets. Therefore, even if the manufacturers change the prices or time for manufacturing or quality as part of the maintenance, the same model can be used to predict impact on the customer preferred optimization parameter as the model is trained accordingly. Table 2 below is a non-exhaustive example for the customer preferred optimization parameter coefficients for each of the customer preferred optimization parameter for each component.

Customer preferred optimization parameter Coefficient for Wheel Configuration Coefficient for Wheel Diameter Coefficient for Rim material Coefficient for Axle distance Coefficient for Suspension System
Price M1a M1b M1c M1d M1n
Manufacturing Time Series M2a M2b M2c M2d M2n
Quality M3a M3b M3c M3d M3n
Manufacturing time parallel M4a M4b M4c M4d M4n

Table 2

[0020] Method step 204 comprises executing the AI model (11) on customer customizations received from the user interface (20) for the product to predict impact on the customer preferred optimization parameter of the product. The AI model (11) runs regression algorithms to predict impact on the customer preferred optimization parameter of the product. Figure 3 depicts the user interface (20) displaying customer preferred optimization parameter for the product based on customer customizations.

[0021] A person skilled in the art will appreciate that while these method steps describes only a series of steps to accomplish the objectives, these methodologies may be implemented with modification to the system (100). This idea to develop a system (100) to predict impact on customer preferred optimization parameter of a product using the training method thereof predicts the kind of impact the given combination might have for the Preferred Optimization Parameter. The customers/users get real-time impact visibility of the combination of the product that they choose. The model also helps in achieving the set of default combinations that the customer might be interested in based on the customer’s preferred optimization parameter that they have selected. Overall, the solution enhances the customer experience while placing order for a complex product by being transparent and giving upfront information on the how the choices of product components impacts in terms of Price, delivery time, quality or any other customer preferred optimization parameter.

[0022] It must be understood that the embodiments explained in the above detailed description are only illustrative and do not limit the scope of this invention. Any modification to the system (100) to predict impact of the customer preferred optimization parameter of a product based on customer customizations and training method thereof are envisaged and form a part of this invention. The scope of this invention is limited only by the claims.

, Claims:We Claim:
1. A system (100) to predict impact on customer preferred optimization parameter for a product based on customer customizations, said system (100) comprising a user interface (20) in communication with a processor (10), characterized in that system (100):
the processor (10) configured to:
retrieve a dataset comprising all possible combinations for a plurality of components of the product and the associated impact of the customer preferred optimization parameter for the product for said combination;
apply feature engineering techniques on the dataset to determine a magnitude of impact, each component has on the customer preferred optimization parameter of the product;
an AI model (11) configured to learn coefficients of each component for predicting the impact on customer preferred optimization parameter using the dataset, the determined magnitude of impact and the values of each component of the plurality of components;
the user interface (20) configured to receive customer customizations for the product from the user;
the processor (10) configured to calculate the customer preferred optimization parameter of user customized product by executing the AI model (11);
the user interface (20) configured to display impact of the customer preferred optimization parameter of the product based on customer customizations.

2. The system (100) to predict impact on customer preferred optimization parameter for a product as claimed in claim 1, wherein customer preferred optimization parameter refers to production time, price, quality of the product.

3. The system (100) to predict impact on customer preferred optimization parameter for a product as claimed in claim 1, wherein feature engineering techniques use either one or more from the group of T-test, Z-test, Pearson coefficient , Cohen values.

4. The system (100) to predict impact on customer preferred optimization parameter for a product as claimed in claim 1, wherein the AI model (11) runs regression algorithms to predict impact on customer preferred optimization parameter of the product.

5. A method (200) of training an AI model (11) to impact on predict customer preferred optimization parameter for a product based on customer customizations, said AI model (11) in communication with a processor (10) of a system (100), the system (100) further comprising a user interface (20) in communication with the processor (10), the method comprising:
retrieving (201) a dataset comprising all possible combinations for a plurality of components of the product and the associated impact of the customer preferred optimization parameter for the product for said combination;
applying (202) feature engineering techniques on the dataset by means of the processor (10) to determine a magnitude of impact each component has on each impact of the customer preferred optimization parameter of the product;
learning (203) coefficients of each component for predicting the impact on customer preferred optimization parameter using the dataset, the determined magnitude of impact and the values of each component of the plurality of components by means of the AI model (11);
executing (204) the AI model (11) on customer customizations received from the user interface (20) for the product to predict impact of the customer preferred optimization parameter of the product.

6. The method (200) of training an AI model (11) as claimed in claim 5, wherein customer preferred optimization parameter refers to production time, price, quality of the product.

7. The method (200) of training an AI model (11) as claimed in claim 5, wherein the feature engineering techniques use either one or more from the group of T-test, Z-test, Pearson coefficient , Cohen values.

8. The method (200) of training an AI model (11) as claimed in claim 5, wherein the AI model (11) runs regression algorithms to predict impact on customer preferred optimization parameter of the product.