Personalized insurance product recommendation method and system based on AI

Abstract

The invention discloses an AI-based personalized insurance product recommendation method and system, and mainly relates to the technical field of application of artificial intelligence in insurance products. Comprising the following steps: collecting basic information and behavior data of a user, and preprocessing the collected basic information and behavior data; dividing the users into new users and old users according to a preprocessing result, and calculating an insurance product recommendation result according to the preprocessed basic information and behavior data of the users for the new users; for old users, analyzing product feature tags selected by the users, and recommending similar products through semantic analysis and feature matching; and performing user behavior sequence analysis on all users through the deep learning model, and predicting potential demands of the users. The method has the advantages that accurate mining and intelligent matching of complex data are achieved, meanwhile, the cold start problem is solved, and efficient recommendation is achieved.

Description

Technical Field

[0001] The present invention relates to the technical field of application of artificial intelligence in insurance products, and specifically to a method and system for recommending personalized insurance products based on AI. Background Art

[0002] As the insurance industry is accelerating its digital transformation, deep learning, big data technology and computer hardware are developing rapidly, AI technology has become an important means to improve efficiency and optimize services. The diversification of insurance products and the growing demand for personalized insurance products from consumers, AI technology can achieve highly personalized insurance plan accurate recommendations, provide a more intelligent and convenient service experience, enhance customer satisfaction and signing rate, and develop personalized recommendation systems based on customer personal data and preferences, which can significantly improve product matching, promote product and service upgrades, and form a competitive advantage.

[0003] Therefore, there is an urgent need for an AI-based personalized insurance product recommendation method and system to solve the above problems. Summary of the invention

[0004] The purpose of the present invention is to provide an AI-based personalized insurance product recommendation method and system, which can realize accurate mining and intelligent matching of complex data, solve the cold start problem, and achieve efficient recommendation.

[0005] In order to achieve the above object, the present invention is implemented through the following technical solutions:

[0006] On the one hand, a personalized insurance product recommendation method based on AI is provided, comprising the following steps:

[0007] S1: Collect basic information and behavior data of users, and pre-process the collected basic information and behavior data;

[0008] S2: Based on the preprocessing results, users are divided into new users and old users. For new users, the preprocessed user basic information and behavior data are used to calculate the insurance product recommendation results; for old users, the product feature tags selected by the user are analyzed, and similar products are recommended through semantic analysis and feature matching;

[0009] S3: Analyze the user behavior sequences of all users through deep learning models to predict their potential needs.

[0010] Preferably, in step S1, the basic information of the user includes: user age, gender and occupation information, and the behavior data includes: browsing, clicking and purchasing record information.

[0011] Preferably, in step S1, the collected basic information and behavior data are preprocessed, specifically: using Python's Spark to clean the data, process missing values ​​and outliers, and perform feature extraction and dimensionality reduction.

[0012] Preferably, in step S2, the specific process of calculating the insurance product recommendation result for the new user is:

[0013] The user's behavior data and product information are represented as vectors, and the dot product of the two vectors is calculated. The dot product is divided by the product of the two vector moduli to obtain the cosine value. The similarity is calculated by the cosine value, which is expressed as:

[0014]

[0015] Where A·B is the dot product of vector A and vector B, ‖A‖ and ‖B‖ are the moduli of vector A and vector B respectively;

[0016] Predict the target user's interest in unrated items based on the preferences of similar users or similar items, and generate a recommendation list.

[0017] Preferably, in step S2, the specific process of recommending similar products to old users through semantic analysis and feature matching is as follows:

[0018] Identify entities and relationships between entities in user information and behavior data, and analyze sentiment trends in behavior data, including positive and negative sentiment trends;

[0019] Identify semantic roles and events in user information and behavior data, and calculate semantic similarity;

[0020] Use natural speech processing technology to extract relevant features from user information and behavior data and establish a collaborative filtering recommendation model;

[0021] Use the collected user information and behavior data to train the recommendation model;

[0022] Based on the similarity calculation results, a recommendation list is generated, listing other products that are most similar to the user input or the current product.

[0023] Preferably, the step S3 predicts the potential needs of the user, specifically:

[0024] The user's age, gender, occupation information, and browsing, clicking, and purchasing record information collected in step S1 are used as extracted features;

[0025] Use TF-IDF algorithm to extract key features from text;

[0026] A recurrent neural network model is established and trained by collecting user information and behavior data.

[0027] On the other hand, a system based on the above AI-based personalized insurance product recommendation method is provided, comprising:

[0028] The data collection layer is used to collect basic information and behavior data of users;

[0029] The data processing layer is used to: clean, feature engineer and format data;

[0030] Recommendation engine model, used to: combine user data and product information and use AI recommendation algorithm to generate recommendation lists;

[0031] The front-end display module is used to: display recommended content through the user interface;

[0032] The backend management module is used to manage products, user feedback, and model updates and maintenance.

[0033] Compared with the prior art, the beneficial effects of the present invention are:

[0034] 1. Compared with traditional technologies, the system can not only extract users' explicit needs, but also discover users' potential needs through machine learning models, recommend highly suitable products to users, and reduce human judgment errors and data bias;

[0035] 2. The hybrid recommendation method provides reasonable recommendations in the case of user cold start, and can efficiently recommend accurate products even when data is scarce, which is a function that is difficult to achieve with traditional recommendation systems;

[0036] 3. Compared with static recommendation methods, reinforcement learning can adjust recommendation strategies in real time based on user clicks, dwell time, purchases and other behavioral feedback to ensure the timeliness and relevance of recommended content;

[0037] 4. Compared with traditional centralized data processing methods, the distributed architecture achieves more efficient processing speed and scalability, allowing the system to maintain efficient recommendation performance when the user scale expands rapidly. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] Figure 1 is a flow chart of the method of the present invention;

[0039] Figure 2 It is a schematic diagram of the system structure of the present invention;

[0040] Figure 3 It is the system control logic flow chart of the present invention. DETAILED DESCRIPTION

[0041] The present invention will be further described below in conjunction with specific embodiments. It should be understood that these embodiments are only used to illustrate the present invention and are not intended to limit the scope of the present invention. In addition, it should be understood that after reading the content taught by the present invention, those skilled in the art can make various changes or modifications to the present invention, and these equivalent forms fall within the scope limited by the application equally.

[0042] In the present invention, terms such as "upper", "lower", "left", "right", "front", "back", "vertical", "horizontal", "side", "bottom", etc. indicate directions or positional relationships based on the directions or positional relationships shown in the accompanying drawings. They are relational words determined only for the convenience of describing the structural relationships of the various parts or elements of the present invention, and do not specifically refer to any part or element in the present invention and should not be understood as limitations on the present invention.

[0043] In the present invention, terms such as "fixed connection", "connected", "connection", etc. should be understood in a broad sense, indicating that it can be fixedly connected, integrally connected or detachably connected; it can be directly connected or indirectly connected through an intermediate medium. Relevant scientific research or technical personnel in this field can determine the specific meanings of the above terms in the present invention according to specific circumstances, and they should not be understood as limiting the present invention.

[0044] Example:

[0045] like Figure 1 As shown, this embodiment provides an AI-based personalized insurance product recommendation method, comprising the following steps:

[0046] S1: Collect basic information and behavior data of users, and pre-process the collected basic information and behavior data;

[0047] S2: Based on the preprocessing results, users are divided into new users and old users. For new users, the preprocessed user basic information and behavior data are used to calculate the insurance product recommendation results; for old users, the product feature tags selected by the user are analyzed, and similar products are recommended through semantic analysis and feature matching;

[0048] S3: Analyze the user behavior sequences of all users through deep learning models to predict their potential needs.

[0049] Among them, in the step S1, the basic information of the user includes: user age, gender and occupation information, and the behavior data includes: browsing, clicking and purchase record information. The collected basic information and behavior data are preprocessed, specifically: using Python's Spark to clean the data, process missing values ​​and outliers, and perform feature extraction and dimensionality reduction.

[0050] In step S2, the specific process of calculating the insurance product recommendation result for the new user is as follows:

[0051] The user's behavior data and product information are represented as vectors, and the dot product of the two vectors is calculated. The dot product is divided by the product of the two vector moduli to obtain the cosine value. The similarity is calculated by the cosine value, which is expressed as:

[0052]

[0053] Where A·B is the dot product of vector A and vector B, ‖A‖ and ‖B‖ are the moduli of vector A and vector B respectively;

[0054] Predict the target user's interest in unrated items based on the preferences of similar users or similar items, and generate a recommendation list.

[0055] In step S2, the specific process of recommending similar products to old users through semantic analysis and feature matching is as follows:

[0056] Identify entities and relationships between entities in user information and behavior data, and analyze sentiment trends in behavior data, including positive and negative sentiment trends;

[0057] Identify semantic roles and events in user information and behavior data, and calculate semantic similarity;

[0058] Use natural speech processing technology to extract relevant features from user information and behavior data and establish a collaborative filtering recommendation model;

[0059] Use the collected user information and behavior data to train the recommendation model;

[0060] Based on the similarity calculation results, a recommendation list is generated, listing other products that are most similar to the user input or the current product.

[0061] The step S3 predicts the potential needs of users, specifically:

[0062] The user's age, gender, occupation information, and browsing, clicking, and purchasing record information collected in step S1 are used as extracted features;

[0063] Use TF-IDF algorithm to extract key features from text;

[0064] A recurrent neural network model is established and trained by collecting user information and behavior data.

[0065] like Figure 2 As shown, this embodiment also provides an AI-based personalized insurance product recommendation system, including:

[0066] The data collection layer is used to collect basic information and behavior data of users;

[0067] The data processing layer is used to: clean, feature engineer and format data;

[0068] Recommendation engine model, used to: combine user data and product information and use AI recommendation algorithm to generate recommendation lists;

[0069] The front-end display module is used to: display recommended content through the user interface;

[0070] The backend management module is used to manage products, user feedback, and model updates and maintenance.

[0071] Its control logic is as follows Figure 3 As shown in the figure, user access and data collection: users access the system through the front-end interface, enter basic information and start browsing insurance products. Data transmission and preprocessing: user data is transmitted to the data server via the Web server, and the data processing layer cleans it. Recommendation engine operation: the recommendation engine receives processed user data and generates a recommendation list through collaborative filtering, deep learning models, etc. Cache and response: the recommendation results are stored in the Redis cache server to reduce the computing overhead during subsequent access. Front-end display and interaction: the front-end displays the recommendation results to the user, and the system adjusts the recommendation strategy based on the user's click feedback.

[0072] The above is a specific description of the preferred implementation of the present invention, but the present invention is not limited to the described embodiments. Those skilled in the art may make various equivalent modifications or substitutions without violating the spirit of the present invention. These equivalent modifications or substitutions are all included in the scope defined by the claims of this application.

Claims

1. A personalized insurance product recommendation method based on AI, characterized in that: The following steps are involved: S1: Collect basic information and behavior data of users, and pre-process the collected basic information and behavior data; S2: Based on the preprocessing results, users are divided into new users and old users. For new users, the preprocessed user basic information and behavior data are used to calculate the insurance product recommendation results; for old users, the product feature tags selected by the user are analyzed, and similar products are recommended through semantic analysis and feature matching; S3: Analyze the user behavior sequences of all users through deep learning models to predict their potential needs.

2. The AI-based personalized insurance product recommendation method according to claim 1, characterized in that: In the step S1, the basic information of the user includes: user age, gender and occupation information, and the behavior data includes: browsing, clicking and purchasing record information.

3. The AI-based personalized insurance product recommendation method according to claim 2, characterized in that: In step S1, the collected basic information and behavior data are preprocessed, specifically: using Python's Spark to clean the data, process missing values ​​and outliers, and perform feature extraction and dimensionality reduction.

4. The AI-based personalized insurance product recommendation method according to claim 1, characterized in that: In step S2, the specific process of calculating the insurance product recommendation result for the new user is as follows: The user's behavior data and product information are represented as vectors, and the dot product of the two vectors is calculated. The dot product is divided by the product of the two vector moduli to obtain the cosine value. The similarity is calculated by the cosine value, which is expressed as: Where A·B is the dot product of vector A and vector B, ‖A‖ and ‖B‖ are the moduli of vector A and vector B respectively; Predict the target user's interest in unrated items based on the preferences of similar users or similar items, and generate a recommendation list.

5. The AI-based personalized insurance product recommendation method according to claim 1, characterized in that: In step S2, the specific process of recommending similar products to old users through semantic analysis and feature matching is as follows: Identify entities and relationships between entities in user information and behavior data, and analyze sentiment trends in behavior data, including positive and negative sentiment trends; Identify semantic roles and events in user information and behavior data, and calculate semantic similarity; Use natural speech processing technology to extract relevant features from user information and behavior data and establish a collaborative filtering recommendation model; Use the collected user information and behavior data to train the recommendation model; Based on the similarity calculation results, a recommendation list is generated, listing other products that are most similar to the user input or the current product.

6. The AI-based personalized insurance product recommendation method according to claim 1, characterized in that: The step S3 predicts the potential needs of users, specifically: The user's age, gender, occupation information, and browsing, clicking, and purchasing record information collected in step S1 are used as extracted features; Use TF-IDF algorithm to extract key features from text; A recurrent neural network model is established and trained by collecting user information and behavior data.

7. A system based on the AI-based personalized insurance product recommendation method according to claim 1, characterized in that: include: The data collection layer is used to collect basic information and behavior data of users; The data processing layer is used to: clean, feature engineer and format data; Recommendation engine model, used to: combine user data and product information and use AI recommendation algorithm to generate recommendation lists; The front-end display module is used to: display recommended content through the user interface; The backend management module is used to manage products, user feedback, and model updates and maintenance.

Images