Insurance business personalized recommendation method based on artificial intelligence

By analyzing user behavior and purchasing historical data, building user portraits to recommend personalized insurance business, it solves the problem that traditional insurance business models are difficult to meet personalized needs, improves recommendation accuracy and marketing efficiency, and reduces costs.

CN119048251AInactive Publication Date: 2024-11-29FULL CHAIN DATA TECHNOLOGY (GUANGZHOU) CO LTD
View PDF 0 Cites 2 Cited by

Patent Information

Application Number
CN202411248480.3
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-09-06
Publication Date
2024-11-29
Estimated Expiration
Not applicable · inactive patent

AI Technical Summary

Technical Problem

The traditional insurance business model relies on manual sales and standardized product design, which is difficult to meet users' personalized needs, resulting in inaccurate recommendations for insurance business.

Method used

By obtaining user behavior data and insurance business purchase historical data, analyzing and processing is used to calculate the user insurance business purchase intention index, and to build user portraits, including interest portraits and situational portraits, and personalized insurance business recommendations are carried out based on these portraits.

Benefits of technology

It improves marketing efficiency, reduces operating costs, provides insurance services that are closer to customer needs, and ensures that every customer receives a professional protection plan.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119048251A_ABST
    Figure CN119048251A_ABST
Patent Text Reader

Abstract

The invention relates to the technical field of insurance business marketing, in particular to an insurance business personalized recommendation method based on artificial intelligence. The method comprises the steps that user behavior data and insurance service purchase historical data are obtained, the user behavior data comprise browsing records of a user on an insurance APP, the behavior data and the insurance service purchase historical data are analyzed and processed, and an insurance service purchase intention index of the user is obtained; and constructing a user portrait according to the insurance business purchase intention index of the user, and performing personalized insurance business recommendation to the user based on the user portrait. According to the invention, the insurance business marketing efficiency is improved, the operation cost is reduced, and the insurance business can be accurately recommended to the user.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the technical field of insurance business marketing, and in particular to a personalized insurance business recommendation method based on artificial intelligence. Background Art

[0002] In today's digital age, with the rapid development of big data, cloud computing and artificial intelligence technologies, all industries are undergoing unprecedented changes, and the insurance industry is no exception. The application of artificial intelligence algorithms, especially machine learning and deep learning technologies, makes it possible to extract valuable information from massive data. Through training models, the system can automatically identify customers' potential needs, risk tolerance and willingness to buy, and then tailor insurance product recommendations for each customer. For example, collaborative filtering algorithms can be used to discover purchasing preferences among customer groups with similar characteristics; and models based on deep learning can more deeply understand unstructured data such as text and images, further enriching customer portraits.

[0003] Traditional insurance business models often rely on manual sales, standardized product design and a wide-ranging marketing strategy. This approach is not only inefficient and unable to meet consumers' growing personalized needs, but may also lead to waste of resources and reduced customer satisfaction.

[0004] Therefore, personalized recommendation methods for insurance business based on artificial intelligence have emerged and become one of the key technologies to promote the transformation and upgrading of the insurance industry. Summary of the invention

[0005] The purpose of the present invention is to provide a personalized insurance business recommendation method based on artificial intelligence: to solve the technical problem that the traditional insurance business model is difficult to meet the personalized needs of users due to its reliance on manual sales, standardized product design and wide-ranging marketing strategies, which in turn leads to inaccurate insurance business recommendations.

[0006] The purpose of the present invention can be achieved through the following technical solutions:

[0007] An insurance business personalized recommendation method based on artificial intelligence, the method comprising:

[0008] Obtain user behavior data and insurance business purchase history data, where user behavior data includes user browsing history on insurance apps;

[0009] Analyze and process behavioral data and insurance business purchase history data to obtain the user's insurance business purchase intention index;

[0010] Determine whether the user's insurance business purchase intention index meets the requirements for obtaining user portrait data, and if so, obtain the user portrait data;

[0011] Build user portraits based on user portrait data, where user portraits include interest portraits and situation portraits;

[0012] Make personalized insurance recommendations to users based on user portraits.

[0013] Furthermore, the behavior data and insurance business purchase history data are analyzed and processed to obtain the user's insurance business purchase intention index, which specifically includes the following process:

[0014] Based on the behavior data, the browsing history of the user on the insurance APP is obtained. Based on the browsing history, the dwell time T of the user browsing the corresponding page of the insurance product, the number of times the user clicks on the corresponding page of the insurance product S, and the number of times the user searches for the insurance product N are obtained;

[0015] Based on the insurance business purchase history data, determine whether the user has purchased insurance business. If so, the user's insurance business purchase intention coefficient R is 1; if not, the user's insurance business purchase intention coefficient R is 0;

[0016] Substitute the stay time T, the number of times S, the number of searches N and the purchase intention coefficient R into the association formula to calculate the user's insurance business purchase intention index GMS. The association formula is as follows:

[0017]

[0018] Among them, γ, θ, and β are weight factors, γ is 0.3, θ is 0.4, and β is 0.3; Y is the number of years of insurance purchased by the user in the purchase history data.

[0019] Furthermore, judging whether the user's insurance business purchase intention index meets the requirements for obtaining user portrait data specifically includes the following process:

[0020] Load the user's insurance business purchase intention index threshold, and determine whether the user's insurance business purchase intention index exceeds the user's insurance business purchase intention index threshold. If so, the requirements for obtaining user portrait data are met; if not, the requirements for obtaining user portrait data are not met.

[0021] Furthermore, constructing a user profile based on the user profile data includes constructing an interest profile based on the user profile data:

[0022] Based on the user portrait data, the texts about insurance business posted by users on social media are obtained, wherein the texts about insurance business are crawled by using crawler technology;

[0023] Based on the LDA topic model, the text is subject-mined. The topic used to characterize the user's interest portrait is determined by calculating the perplexity of each topic, and the topic feature words corresponding to the topic are extracted.

[0024] The topic feature words are used as interest preference feature tags, and the user's insurance business interest profile is generated based on the interest preference feature tags.

[0025] Furthermore, the topic mining of the text is carried out based on the LDA topic model. The topic used to characterize the user's interest portrait is determined by calculating the perplexity of each topic, and the topic feature words corresponding to the topic are extracted. The specific process includes the following:

[0026] Generate topics for text based on the LDA topic model:

[0027] (1) Sampling from the prior Dirichlet distribution α generates the topic multinomial distribution α of document m m ;

[0028] (2) From the multinomial distribution of topics α m The topic Z of the nth word in document m is sampled m,n ;

[0029] (3) Generate topic Z by sampling from the Dirichlet distribution δ of the vocabulary m,n The multinomial distribution of words θ zmn ;

[0030] (4) From the multinomial distribution of words θ zmn Sampling generated words W m,n ;

[0031] (5) The word W m,n as the subject of document m;

[0032] The topics used to characterize user interest portraits are determined by calculating the perplexity of each topic:

[0033]

[0034] Per(D) is the perplexity of each topic. The lower the perplexity, the more accurate the topic is in describing the user’s interest profile. D is the dataset consisting of all topics. K is the number of documents corresponding to the topic. w d represents the words that appear in document d, p(w d ) represents the word w in each document d The probability of occurrence, N d Represents the number of words in document d;

[0035] Sort the topics in the dataset D according to the degree of perplexity, select the k lowest-ranked topics to characterize the user's interest portrait, and extract the topic feature words corresponding to the topic.

[0036] Furthermore, constructing a user portrait based on the user portrait data includes constructing a situational portrait based on the user portrait data:

[0037] User portrait data includes the user's insurance business purchase history data, based on which the user's insurance business purchase history data is used to obtain the user's transaction data in the insurance business, including policy information, purchase records, and claims records;

[0038] Extract key features based on transaction data, including age, gender, occupation, income level, purchase preference, and risk tolerance;

[0039] The key features are used as contextual feature labels, and the user's insurance business contextual profile is generated based on the contextual feature labels.

[0040] Furthermore, personalized insurance business recommendations to users based on user portraits include:

[0041] According to the user situational preferences and situational characteristics reflected in the user situational portrait, an association analysis is performed on the user situational preferences and situational characteristics and the insurance business to be selected by the user, the association rules between the user situational preferences and insurance business are mined, and an association rule library is constructed. The situational similarity is calculated based on the situational characteristics and preferences of the target user and the context of the association rules, and the rules with higher similarity are selected. Recommendations are made to the target user based on the rules with higher similarity, and a set A of recommended insurance business items is obtained, and recommendations are made to the user based on set A.

[0042] Furthermore, personalized insurance business recommendations to users based on user portraits include:

[0043] Based on the user interest tags reflected in the user interest portrait, the interest similarity between users is calculated by calculating the similarity of user interest feature words. The user with the highest interest similarity to the target user is recorded as the nearest neighbor user, and the insurance business recommendation set of the nearest neighbor user is recommended to the target user.

[0044] Compared with the existing solutions, the present invention achieves the following beneficial effects:

[0045] The present invention obtains the user's insurance purchase intention index by analyzing and processing the behavior data and insurance purchase history data, and constructs a user profile based on the user's insurance purchase intention index, and makes personalized insurance recommendations to the user based on the user profile. The present invention not only improves marketing efficiency and reduces operating costs, but also makes insurance services more in line with customer needs, providing each customer with a more intimate and professional protection plan. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in the present invention. For ordinary technicians in this field, other drawings can also be obtained based on these drawings.

[0047] Figure 1 This is a workflow diagram of a personalized insurance business recommendation method based on artificial intelligence according to an embodiment of the present invention;

[0048] Figure 2 This is a workflow diagram of another method for personalized recommendation of insurance services based on artificial intelligence according to an embodiment of the present invention;

[0049] Figure 3 This is a workflow diagram of another method for personalized recommendation of insurance business based on artificial intelligence according to an embodiment of the present invention. DETAILED DESCRIPTION

[0050] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0051] In addition, the described features, structures or characteristics may be combined in one or more example embodiments in any suitable manner. In the following description, many specific details are provided to provide a full understanding of the example embodiments of the present disclosure. However, those skilled in the art will appreciate that the technical solutions of the present disclosure may be practiced while omitting one or more of the specific details, or other methods, components, steps, etc. may be adopted. In other cases, well-known structures, methods, implementations or operations are not shown or described in detail to avoid obscuring various aspects of the present disclosure.

[0052] This embodiment provides a personalized insurance business recommendation method based on artificial intelligence. Figure 1 is a workflow diagram of an insurance business personalized recommendation method based on artificial intelligence according to an embodiment of the present invention, such as Figure 1 As shown, the method comprises the following steps:

[0053] Step S101: Obtain user behavior data and insurance business purchase history data, wherein the user behavior data includes the user's browsing history on the insurance APP;

[0054] Step S102: Analyze and process the behavior data and insurance business purchase history data to obtain the user's insurance business purchase intention index;

[0055] Step S103: determine whether the user's insurance business purchase intention index meets the requirements for obtaining user portrait data, if so, proceed to step S104, if not, proceed to step S107;

[0056] Step S104: Obtain user portrait data;

[0057] Step S105: constructing a user portrait based on the user portrait data, wherein the user portrait includes an interest portrait and a situation portrait;

[0058] Step S106: Make personalized insurance recommendations to users based on their profiles.

[0059] Step S107: Refuse to obtain user portrait data.

[0060] In summary, the present invention obtains user behavior data and insurance business purchase history data, wherein the user behavior data includes the user's browsing history on the insurance APP, analyzes and processes the behavior data and insurance business purchase history data, obtains the user's insurance business purchase intention index, and constructs a user portrait based on the user's insurance business purchase intention index, and makes personalized insurance business recommendations to the user based on the user portrait. The present invention not only improves marketing efficiency and reduces operating costs, but also can accurately recommend insurance business to users.

[0061] In some embodiments, Figure 2 is a workflow diagram of another method for personalized insurance business recommendation based on artificial intelligence according to an embodiment of the present invention, such as Figure 2 As shown, analyzing and processing the behavior data and insurance business purchase history data to obtain the user's insurance business purchase intention index specifically includes the following steps:

[0062] Step S201: Obtain the browsing history of the user on the insurance APP based on the behavior data, and obtain the dwell time T of the user browsing the page corresponding to the insurance product, the number of times the user clicks the page corresponding to the insurance product S, and the number of times the user searches for the insurance product N based on the browsing history;

[0063] Step S202: judging whether the user has purchased insurance services based on the insurance business purchase history data, if so, the user's insurance business purchase intention coefficient R is 1, if not, the user's insurance business purchase intention coefficient R is 0;

[0064] Step S203: Substitute the stay time T, the number of times S, the number of searches N and the purchase intention coefficient R into the correlation formula to calculate the user's insurance business purchase intention index GMS. The correlation formula is as follows:

[0065]

[0066] Among them, γ, θ, and β are weight factors, γ is 0.3, θ is 0.4, and β is 0.3; Y is the number of years of insurance purchased by the user in the purchase history data.

[0067] In some embodiments, in step S103, determining whether the user's insurance business purchase intention index meets the requirements for obtaining user portrait data specifically includes the following process:

[0068] Load the user's insurance business purchase intention index threshold, and determine whether the user's insurance business purchase intention index exceeds the user's insurance business purchase intention index threshold. If so, the requirements for obtaining user portrait data are met; if not, the requirements for obtaining user portrait data are not met.

[0069] In some embodiments, Figure 3 is a workflow diagram of another method for personalized insurance business recommendation based on artificial intelligence according to an embodiment of the present invention, such as Figure 3 As shown, building a user profile based on user profile data includes building an interest profile based on the user profile data:

[0070] Step S301: obtaining texts about insurance business posted by users on social media based on user portrait data, wherein the texts about insurance business are crawled by using crawler technology;

[0071] Step S302: Perform topic mining on the text based on the LDA topic model, determine the topic used to characterize the user's interest portrait by calculating the perplexity of each topic, and extract the topic feature words corresponding to the topic;

[0072] Step S303: Use the topic feature words as interest preference feature tags, and generate the user's insurance business interest portrait based on the interest preference feature tags.

[0073] It is worth mentioning that LDA (Latent Dirichlet Allocation) is a document generation model that assumes that each document has multiple topics, and each topic has its characteristic words. The topic of each document in the document set is given in the form of probability distribution, and the topic is extracted by analyzing the document, and then the topic is clustered.

[0074] In some embodiments, in step S302, topic mining is performed on the text based on the LDA topic model, the topic used to describe the user's interest portrait is determined by calculating the perplexity of each topic, and the topic feature words corresponding to the topic are extracted, which specifically includes the following process:

[0075] Generate topics for text based on the LDA topic model:

[0076] (1) Sampling from the prior Dirichlet distribution α generates the topic multinomial distribution α of document m m ;

[0077] (2) From the multinomial distribution of topics α m The topic Z of the nth word in document m is sampled m,n ;

[0078] (3) Generate topic Z by sampling from the Dirichlet distribution δ of the vocabulary m,n The multinomial distribution of words θ zmn ;

[0079] (4) From the multinomial distribution of words θ zmn Sampling generated words W m,n ;

[0080] (5) The word W m,n as the subject of document m;

[0081] The topics used to characterize user interest portraits are determined by calculating the perplexity of each topic:

[0082]

[0083] Per(D) is the perplexity of each topic. The lower the perplexity, the more accurate the topic is in describing the user’s interest profile. D is the dataset consisting of all topics. K is the number of documents corresponding to the topic. w d represents the words that appear in document d, p(w d ) represents the word w in each document d The probability of occurrence, N d Represents the number of words in document d;

[0084] Sort the topics in the dataset D according to the degree of perplexity, select the k lowest-ranked topics to characterize the user's interest portrait, and extract the topic feature words corresponding to the topic.

[0085] In some embodiments, constructing a user profile based on the user profile data includes constructing a situational profile based on the user profile data:

[0086] User portrait data includes the user's insurance business purchase history data, based on which the user's insurance business purchase history data is used to obtain the user's transaction data in the insurance business, including policy information, purchase records, and claims records;

[0087] Key features are extracted based on transaction data, including age, gender, occupation, income level, purchasing preference, and risk tolerance.

[0088] Specifically, the frequency of occurrence of words in the text is calculated based on the Bag of Words (BoW) model to generate a feature vector.

[0089] TF-IDF (TermFrequency-Inverse DocumentFrequency): Considers the importance and general importance of words, improves the bag-of-words model, assigns lower weights to common words, and assigns higher weights to rare words.

[0090] Text Embedding: Use pre-trained models such as Word2Vec, GloVe, or BERT to convert text into high-dimensional vectors that can capture the semantic relationship between words.

[0091] Keyword extraction: Use RAKE (Rapid Automatic Keyword Extraction), TextRank and other algorithms to extract keywords from text.

[0092] Named Entity Recognition (NER): Identify specific entities in text, such as names of people, places, products, etc., which may be important features.

[0093] The key features are used as contextual feature labels, and the user's insurance business contextual profile is generated based on the contextual feature labels.

[0094] In some embodiments, recommending personalized insurance services to users based on user portraits includes:

[0095] Step 1: Association rule mining

[0096] The Apriori algorithm is used to mine association rules. Based on the context labels such as age, gender, occupation, income level, purchase preference, and risk tolerance in the user context portrait, the context information of all users is discretized and normalized as the antecedent and the insurance business as the posterior. The multidimensional association rules between the tourist context characteristics and the insurance business are mined, such as: age ∧ gender ∧ occupation ∧ income level ∧ purchase preference ∧ risk tolerance The insurance business mines strong association rules by setting the minimum support (min-sup) and minimum confidence (min-conf) thresholds. When the support and confidence of an association rule are greater than the minimum support and minimum confidence, it is determined to be a strong association rule.

[0097] Step 2: Calculate the context similarity

[0098] First, the local similarity between the target user and each context of the strong association rule is calculated; secondly, the distance between the target user and each context of the strong association rule is calculated; then, the distance between the two is calculated using the context weight; finally, the similarity between the two is calculated and the Euclidean distance is used to calculate the similarity.

[0099] Assume that the context set of the target user is Y = {X1, X2, ...X j}, the context value in scenario b is X j ={x b1 , x b2 ,...x bj}, the context set of strong association rules is H = {Z1, Z2, ... Z j}, the situation value in situation b is Z j ={z b1 , z b2 ,...z bj}, if the target user has the same characteristics as the strong association rule in the same context, then the local similarity sim(X j , Z j ) is 1, otherwise the similarity sim(X j , Z j ) is 0.

[0100] The distance DS between the target user and the strong association rule in context b is calculated as follows:

[0101] DS(X j , Z j )=1-sim(X j , Z j );

[0102] The distance between the target user and the strong association rule is as follows, where W j is the weight of the context, which is calculated based on the frequency of different contexts in all strong association rules. j ∈(0,1):

[0103]

[0104] The similarity between the target user and the strong association rule SIM(X j , Z j )for:

[0105]

[0106] Step 3: Generate recommendations

[0107] By calculating the similarity between the contexts of the target user and all strong association rules, these strong association rules are arranged in descending order according to the similarity to form a set of strong association rules with high context similarity to the target user. The attractions in the latter part of the strong association rules are recommended to the target user according to the similarity from high to low to obtain the recommendation set A.

[0108] In some embodiments, recommending personalized insurance services to users based on user portraits includes:

[0109] Based on the user interest tags reflected in the user interest portrait, the interest similarity between users is calculated by calculating the similarity of user interest feature words. The user with the highest interest similarity to the target user is recorded as the nearest neighbor user, and the insurance business recommendation set of the nearest neighbor user is recommended to the target user.

[0110] In order to calculate the similarity of interests between users, a variety of methods can be used, such as cosine similarity, Jaccard similarity, etc. Here we take Jaccard similarity as an example because it is suitable for processing set data. Jaccard similarity is defined as the ratio of the size of the intersection of two sets to the size of their union.

[0111] The above embodiments can be implemented in whole or in part by software, hardware, firmware or any other combination. When implemented by software, the above embodiments can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, the process or function described in the embodiment of the present application is generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium, or transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, the computer instructions can be transmitted from one website site, computer, server or data center to another website site, computer, server or data center by wired (e.g., infrared, wireless, microwave, etc.). The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that contains one or more available media sets. The available medium can be a magnetic medium (e.g., a floppy disk, a hard disk, a tape), an optical medium (e.g., a DVD), or a semiconductor medium. The semiconductor medium can be a solid-state hard disk.

[0112] Those of ordinary skill in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.

[0113] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0114] In the several embodiments provided in the present application, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of the units is only some logical function divisions. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0115] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0116] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art who is familiar with the present technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

Claims

1. A personalized insurance business recommendation method based on artificial intelligence, characterized in that: Methods include: Obtain user behavior data and insurance business purchase history data, where user behavior data includes user browsing history on insurance apps; Analyze and process behavioral data and insurance business purchase history data to obtain the user's insurance business purchase intention index; Determine whether the user's insurance business purchase intention index meets the requirements for obtaining user portrait data, and if so, obtain the user portrait data; Build user portraits based on user portrait data, where user portraits include interest portraits and situation portraits; Make personalized insurance recommendations to users based on user portraits.

2. The method for personalized insurance business recommendation based on artificial intelligence according to claim 1, characterized in that: The process of analyzing and processing the behavior data and insurance business purchase history data to obtain the user's insurance business purchase intention index specifically includes the following steps: Based on the behavior data, the browsing history of the user on the insurance APP is obtained. Based on the browsing history, the dwell time T of the user browsing the corresponding page of the insurance product, the number of times the user clicks on the corresponding page of the insurance product S, and the number of times the user searches for the insurance product N are obtained; Based on the insurance business purchase history data, determine whether the user has purchased insurance business. If so, the user's insurance business purchase intention coefficient R is 1; if not, the user's insurance business purchase intention coefficient R is 0; Substitute the stay time T, the number of times S, the number of searches N and the purchase intention coefficient R into the association formula to calculate the user's insurance business purchase intention index GMS. The association formula is as follows: Among them, γ, θ, and β are weight factors, γ is 0.3, θ is 0.4, and β is 0.3; Y is the number of years of insurance purchased by the user in the purchase history data.

3. The method for personalized insurance business recommendation based on artificial intelligence according to claim 1, characterized in that: The process of judging whether the user's insurance business purchase intention index meets the requirements for obtaining user portrait data specifically includes the following steps: Load the user's insurance business purchase intention index threshold, and determine whether the user's insurance business purchase intention index exceeds the user's insurance business purchase intention index threshold. If so, the requirements for obtaining user portrait data are met; if not, the requirements for obtaining user portrait data are not met.

4. The method for personalized insurance business recommendation based on artificial intelligence according to claim 1, characterized in that: Building a user profile based on user profile data includes building an interest profile based on user profile data: Based on the user portrait data, the texts about insurance business posted by users on social media are obtained, wherein the texts about insurance business are crawled by using crawler technology; Based on the LDA topic model, the text is subject-mined. The topic used to characterize the user's interest portrait is determined by calculating the perplexity of each topic, and the topic feature words corresponding to the topic are extracted. The topic feature words are used as interest preference feature tags, and the user's insurance business interest profile is generated based on the interest preference feature tags.

5. The method for personalized insurance business recommendation based on artificial intelligence according to claim 4, characterized in that: Based on the LDA topic model, the topic mining of the text is carried out. The topic used to characterize the user's interest portrait is determined by calculating the perplexity of each topic, and the topic feature words corresponding to the topic are extracted. The specific process includes the following: Generate topics for text based on the LDA topic model: (1) Sampling from the prior Dirichlet distribution α generates the topic multinomial distribution α of document m m ; (2) From the multinomial distribution of topics α m The topic Z of the nth word in document m is sampled m,n ; (3) Generate topic Z by sampling from the Dirichlet distribution δ of the vocabulary m,n The multinomial distribution of words θ zmn ; (4) From the multinomial distribution of words θ zmn Sampling generated words W m,n ; (5) The word W m,n as the subject of document m; The topics used to characterize user interest portraits are determined by calculating the perplexity of each topic: Per(D) is the perplexity of each topic. The lower the perplexity, the more accurate the topic is in describing the user’s interest profile. D is the dataset consisting of all topics. K is the number of documents corresponding to the topic. w d represents the words that appear in document d, p(w d ) represents the word w in each document d The probability of occurrence, N d Represents the number of words in document d; Sort the topics in the dataset D according to the degree of perplexity, select the k lowest-ranked topics to characterize the user's interest portrait, and extract the topic feature words corresponding to the topic.

6. The method for personalized insurance business recommendation based on artificial intelligence according to claim 1, characterized in that: Building a user portrait based on user portrait data includes building a situational portrait based on user portrait data: User portrait data includes the user's insurance business purchase history data, based on which the user's insurance business purchase history data is used to obtain the user's transaction data in the insurance business, including policy information, purchase records, and claims records; Extract key features based on transaction data, including age, gender, occupation, income level, purchase preference, and risk tolerance; The key features are used as contextual feature labels, and the user's insurance business contextual profile is generated based on the contextual feature labels.

7. The method for personalized insurance business recommendation based on artificial intelligence according to claim 1, characterized in that: Personalized insurance recommendations to users based on user portraits include: According to the user situational preferences and situational characteristics reflected in the user situational portrait, an association analysis is performed on the user situational preferences and situational characteristics and the insurance business to be selected by the user, the association rules between the user situational preferences and insurance business are mined, and an association rule library is constructed. The situational similarity is calculated based on the situational characteristics and preferences of the target user and the context of the association rules, and the rules with higher similarity are selected. Recommendations are made to the target user based on the rules with higher similarity, and a set A of recommended insurance business items is obtained, and recommendations are made to the user based on set A.

8. The method for personalized insurance business recommendation based on artificial intelligence according to claim 1, characterized in that: Personalized insurance recommendations to users based on user portraits include: Based on the user interest tags reflected in the user interest portrait, the interest similarity between users is calculated by calculating the similarity of user interest feature words. The user with the highest interest similarity to the target user is recorded as the nearest neighbor user, and the insurance business recommendation set of the nearest neighbor user is recommended to the target user.

Citation Information

Cited By

  • Data tracking analysis management platform based on digital life insurance enterprise

    CN120106994A

  • A data tracking, analysis and management platform based on digital life insurance companies

    CN120106994B