Car insurance intelligent quotation system based on multi-dimensional data analysis

Abstract

The invention provides a car insurance intelligent quotation system based on multi-dimensional data analysis, and the system comprises the following modules: a front-end configuration management module which carries out the rule configuration at a business front-end configuration interface, and obtains an electronic form; the intelligent analysis module is used for extracting keywords in the electronic form by adopting OCR (Optical Character Recognition) and preprocessing data to obtain data I; and the scheme matching module takes the data I as the input of a transformer model, outputs a scheme with a high matching degree and predicts an insurance conversion rate corresponding to the scheme. Intelligent quotation is performed by fusing multi-source heterogeneous data, a dynamic configuration rule engine and a machine learning algorithm, a vehicle-user-driver protection three-dimensional data association model is constructed, an end-to-end intelligent quotation system from data acquisition, demand analysis to scheme recommendation is realized, and quotation accuracy and customer conversion efficiency are remarkably improved.

Description

Technical Field

[0001] This application relates to the field of insurance quotation service systems, and particularly to a motor vehicle insurance intelligent quotation system based on multi-dimensional data analysis. Background Art

[0002] With the accelerating digital transformation of property insurance business, problems existing in the traditional motor vehicle insurance quotation mode, such as low efficiency, high dependence on manual labor, and insufficient standardization, have become increasingly prominent. Currently, insurance intermediaries need to rely on back-office personnel to manually complete processes such as insurance company selection, insurance type configuration, information entry, and quotation generation in motor vehicle insurance business. The following technical pain points exist: Manual operations are time-consuming and error-prone, making it difficult to meet the high-frequency and multi-scenario customer needs; The configuration of insurance plans lacks intelligence and cannot achieve personalized recommendations based on multi-dimensional dynamic factors (such as vehicle attributes, vehicle owner information, regional characteristics, insurance company policies, etc.); The product data of different insurance companies is scattered, and quotation comparison and plan optimization rely on empirical judgment, lacking systematic algorithm support. In the prior art, most motor vehicle insurance quotation systems adopt rule engines or simple conditional screening mechanisms, and their limitations are as follows: The data dimension is single, and vehicle characteristics (energy type, vehicle model, years of use, etc.), policyholder portraits, and insurance company dynamic strategies (such as commission policy rules, benefit insurance type binding rules) are not deeply integrated; There is a lack of an intelligent decision-making model based on complex factor weights, making it difficult to achieve the priority ranking of insurance company plans and the optimal matching of loss ratio - expense ratio; The efficiency of human-machine collaboration is insufficient, and manual intervention is required in links such as OCR parsing result verification and insurance type combination adjustment. Summary of the Invention

[0003] In view of the above problems, the present invention provides a motor vehicle insurance intelligent quotation system based on multi-dimensional data analysis, which integrates multi-source heterogeneous data, dynamically configures a rule engine and machine learning algorithms for intelligent quotation, constructs a three-dimensional data association model of vehicle-user-insurance company, and realizes an end-to-end intelligent quotation system from data collection, requirement analysis to plan recommendation, significantly improving the quotation accuracy and customer conversion efficiency.

[0004] In a first aspect, the present invention provides a motor vehicle insurance intelligent quotation system based on multi-dimensional data analysis, including the following: A front-end configuration management module, which performs rule configuration on the business front-end configuration interface to obtain an electronic form; wherein, the electronic form includes the basic information of the policyholder, the basic information of the vehicle, expected information, and the dynamic strategy of the insurance company.

[0005] An intelligent parsing module, which extracts keywords from the electronic form using OCR, and after preprocessing the data, obtains Data 1.

[0006] The solution matching module takes the first data as the input of the rule engine model, outputs a solution with a high matching degree, and predicts the insurance conversion rate corresponding to the solution; wherein, the rule engine model uses a transformer model to learn and extract feature values, and predicts a solution with a high conversion rate and a match.

[0007] Further, the OCR extracts keywords in the electronic form and uses a deep network model for text recognition. The specific operations are as follows:

[0008] Construct a bidirectional Conv-LSTM model and train the model with a large amount of data to obtain a recognition model;

[0009] Collect the picture data of the electronic form, perform feature extraction and feature sequence modeling through the trained recognition model, and finally use CTC decoding processing to obtain the recognized text.

[0010] Further, the rule engine model uses a transformer structure to predict matching solutions. The specific operations are as follows:

[0011] Perform feature engineering on the collected solution matching samples to obtain a trainable data set;

[0012] The encoding network of the Transformer uses a pre-trained VGG model, connects a classification task head, and trains the transformer model with the training data set;

[0013] Collect target data as the input of the Transformer model to obtain a matching solution category; wherein, the attention mechanism of the Transformer is:

[0014]

[0015] Further, the solution is configured with three-tier package solutions of low return, medium return, and high return.

[0016] Further, the keyword is configured to be autonomously extracted in the intelligent parsing module and stored as an element in the data.

[0017] Further, the storage method of the element includes a dictionary or XML format.

[0018] In a second aspect, a computer-readable storage medium stores instructions thereon, and when the instructions are executed by a processor, the vehicle insurance intelligent quotation system based on multi-dimensional data analysis as described in any one of the above is implemented.

[0019] Through the above implementation solutions, the following advantages or beneficial effects are achieved:

[0020] (1) Integrate multi-dimensional data features to improve the accuracy of model prediction, thereby enhancing the precision of intelligent quotation;

[0021] (2) Multi-objective optimization strategy to dynamically adapt to the reasonable interests of multiple parties and improve the customer conversion efficiency after quotation;

[0022] (3) End-to-end intelligent quotation prediction, which improves the business handling efficiency and thus saves costs.

[0023] Other features and advantages of this application will be described in the subsequent specification, and some of them will become obvious from the specification or be understood by implementing this application. The objectives and other advantages of this application can be achieved and obtained through the structures specifically pointed out in the written specification, claims, and drawings. Description of the Drawings

[0024] To more clearly illustrate the technical solutions in the embodiments of this application or related technologies, the following will briefly introduce the drawings required for use in the description of the embodiments or related technologies. Obviously, the drawings in the following description are only the embodiments of this application. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on the provided drawings.

[0025] Figure 1 It is a flowchart of a vehicle insurance intelligent quotation system based on multi-dimensional data analysis provided by the present invention. Detailed Embodiments

[0026] To make the objectives, technical solutions, and advantages of this application clearer, the following will clearly and completely describe the technical solutions in the embodiments of this application with reference to the drawings in the embodiments of this application. Obviously, the described embodiments are only part of the embodiments of this application, rather than all of them. Based on the embodiments in this application, all other embodiments obtained by those of ordinary skill in the art without creative efforts belong to the scope of protection of this application.

[0027] Embodiment 1

[0028] As Figure 1 shown, a vehicle insurance intelligent quotation system based on multi-dimensional data analysis includes the following modules:

[0029] A front-end configuration management module configures rules in the business front-end configuration interface to obtain an electronic form. Among them, the electronic form includes the basic information of the applicant, the basic information of the vehicle, the expected information, and the dynamic strategy of the insurance company;

[0030] The described front-end configuration management module must configure rules in the front-end configuration interface to generate an electronic form including the basic information of the applicant, the basic information of the vehicle, the expected information, and the dynamic policies of the insurance company; the basic information of the applicant includes age, gender, and occupation; the basic information of the vehicle includes energy type, vehicle model classification, and service life; the expected information includes the level of risk assumption; the dynamic policies of the insurance company include commission policy rules and benefit insurance type binding rules.

[0031] The intelligent parsing module uses OCR to extract keywords in the electronic form, and after preprocessing the data, obtains Data 1;

[0032] If the data collected in the front-end configuration is in the form of a pdf or a picture, the intelligent parsing module uses OCR to extract the keyword fields in the electronic form, and the extracted keywords are configured to be autonomously extracted in the intelligent parsing module and stored as elements in the data. The storage method of the elements includes dictionary or XML format.

[0033] The example of storing in dictionary form is as follows: For example: { { age: 39; gender: male; occupation: freelancer}; { energy type: hybrid; vehicle model classification: car; service life: 9}; { risk assumption: small}; { commission policy rule: 3; benefit insurance type binding rule: 2}; { whether to insure: yes}; { package plan: Y}}, where the commission policy rule is set to 3 levels, and which level it conforms to, then the corresponding code; the benefit insurance type binding rule is set to 4 levels, and the code that conforms to the level is filled in the corresponding dictionary; in addition, Insurance Company X package combination (compulsory traffic insurance + third-party liability insurance of 2 million + additional out-of-hospital medical expenses of 200,000, total price 4,500 yuan), Insurance Company Y package (same insurance type and insurance amount + three times of road rescue rights, total price 4,800 yuan).

[0034] Adopt the dictionary storage method in the above example to collect a batch of training data for model training.

[0035] When the electronic form is a PDF scan or a picture, the OCR extracts the keywords in the electronic form and uses a deep network model for text recognition. The specific operations are as follows:

[0036] Construct a bidirectional Conv-LSTM model and use a large amount of data to train the model to obtain a recognition model;

[0037] Collect the picture data of the electronic form, perform feature extraction and feature sequence modeling through the trained recognition model, and finally use CTC decoding processing to obtain the recognized text.

[0038] In practical applications, the data in the electronic form is vectorized and preprocessed to obtain Data One; this Data One is used as the input of the model to predict the corresponding insurance plan and the conversion rate of insurance application.

[0039] The plan matching module takes the Data One as the input of the rule engine model, outputs a plan with a high matching degree and predicts the insurance conversion rate corresponding to the plan; wherein, the rule engine model uses the transformer model to learn and extract feature values to predict a plan with a high conversion rate and a good match.

[0040] The rule engine model uses the transformer structure to predict the matching plan, and the specific operation is as follows:

[0041] Perform feature engineering on the collected plan matching samples to obtain a trainable data set;

[0042] The main framework of Transformer uses a pre-trained VGG model, connects the classification task head, and trains the transformer model using the training data set; [[ID=q13]]

[0043] Collect target data as the input of the Transformer model to obtain the matching plan category; wherein, the attention mechanism of Transformer is: ]>

[0044]

[0045] Use a batch of training data collected above to train the rule engine model constructed by the transformer model to obtain a rule engine model after training convergence; and use the Data One as the input of the rule engine model to output the predicted package plan and the corresponding insurance conversion rate.

[0046] Embodiment 2

[0047] A computer-readable storage medium, on which instructions are stored, and when the instructions are executed by a processor, the vehicle insurance intelligent quotation system based on multi-dimensional data analysis as described in any one of the above is implemented.

[0048] Through the above solutions, the following advantages or beneficial effects are achieved:

[0049] (1) Integrate multi-dimensional data features, improve the accuracy of model prediction, and thus improve the accuracy of intelligent quotation;

[0050] (2) The multi-objective optimization strategy dynamically adapts to the reasonable interests of multiple parties and improves the customer conversion efficiency after quotation;

[0051] (3) End-to-end intelligent quotation prediction improves the business handling efficiency and thus saves costs.

[0052] As described above, it is only the specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any changes or substitutions that can be easily thought of by those skilled in the art within the technical scope disclosed by the present invention should be covered within the protection scope of the present invention. Therefore, the protection scope of the present invention shall be subject to the protection scope of the claimed rights.

Claims

1. A vehicle insurance intelligent quotation system based on multi-dimensional data analysis, characterized in that, It includes the following: a front-end configuration management module that configures rules in the business front-end configuration interface to obtain an electronic form; wherein, the electronic form includes basic information of the applicant, basic information of the vehicle, expected information, and the insurance company's dynamic policies; An intelligent parsing module that uses OCR to extract keywords in the electronic form and, after preprocessing the data, obtains Data 1; A solution matching module that takes Data 1 as the input of a rule engine model, outputs a highly matched solution, and predicts the insurance application conversion rate corresponding to the solution; wherein, the rule engine model uses a transformer model to learn and extract feature values and predicts a highly convertible and matched solution.

2. The vehicle insurance intelligent quotation method based on multi-dimensional data analysis according to claim 1, wherein, The OCR extracts keywords in the electronic form and uses a deep network model for text recognition. The specific operations are as follows: Construct a bidirectional Conv-LSTM model and train the model with a large amount of data to obtain a recognition model; Collect image data of the electronic form, perform feature extraction and feature sequence modeling through the trained recognition model, and finally use CTC decoding processing to obtain the recognized text.

3. The intelligent car insurance quotation method based on multi-dimensional data analysis according to claim 1, characterized in that The rule engine model uses a transformer structure to predict a matching solution. The specific operations are as follows: Perform feature engineering on the collected solution matching samples to obtain a trainable data set; The encoding network of the Transformer uses a pre-trained VGG model, connects a classification task head, and trains the transformer model with the training data set; Collect target data as the input of the Transformer model to obtain a matching solution category; wherein, the attention mechanism of the Transformer is:

4. The intelligent car insurance quotation method based on multi-dimensional data analysis according to claim 1, characterized in that, The solution is configured with three-tier package solutions of low return, medium return, and high return.

5. The intelligent car insurance quotation method based on multi-dimensional data analysis according to claim 1, characterized in that The keyword is configured to be autonomously extracted in the intelligent parsing module and stored as an element in the data.

6. The intelligent car insurance quotation method based on multi-dimensional data analysis according to claim 5, characterized in that, The storage method of the element includes a dictionary or XML format.

7. A computer-readable storage medium having instructions stored thereon, characterized in that, When the instruction is executed by the processor, it implements the vehicle insurance intelligent quotation system based on multi-dimensional data analysis as described in any one of claims 1-6.

Images