Personalized service system and method based on big data

Abstract

The invention discloses a personalized service system and method based on big data, relates to the technical field of big data analysis, and solves the problems that when the number of people who use a package is one, a family package is not suitable, and the special situation that a single person uses the package is not considered, so that a recommended package cannot meet the actual demand of a user, and the user experience is poor. And the use experience of the user is influenced. According to the method, basic information, real-time behavior data and historical business data of a plurality of channel users are acquired; constructing a user portrait according to the basic information of the user; determining a transaction service of the user according to the user portrait and the real-time behavior data; analyzing personalized data of the user according to the historical business data of the user; recommending a corresponding service package to the user based on the personalized data of the user; monitoring the real-time flow data of the user; value-added services are recommended according to the real-time traffic of the user; the service can be recommended according to the actual demand of the user, and the satisfaction and loyalty of the user can be improved.

Description

Technical Field

[0001] The present invention belongs to the field of big data analysis, relates to personalized service technology, and specifically is a personalized service system and method based on big data. Background Art

[0002] The services of operators mainly include services such as online card handling, broadband services, and value-added services; operators have moved traditional business hall services to online platforms, making business handling more convenient and fast; users can apply for mobile communication services online through the websites or application programs of operators, saving time and energy; personalized services can provide customized communication services based on users' needs and preferences; by analyzing users' communication habits, usage scenarios, and preferences, operators can recommend suitable packages, data packages, value-added services, etc. for users, thereby improving users' satisfaction and loyalty; personalized services can meet users' unique needs and make users feel valued and concerned; this differentiated service can enhance users' dependence on operators and increase user stickiness.

[0003] The prior art (a patent application with publication number CN118195733A) discloses a method for recommending operator home services; the method includes: obtaining home basic data and service-related data; determining the operator service usage portraits of family members and the main consumption information of operator services according to the home basic data; determining the comparison results of each item and the comprehensive service comparison results according to the service-related data; determining a home recommended service matrix according to the operator communication service usage portraits of each family member, the main consumption information of operator services, and the comparison results of each item, and the home recommended service matrix includes recommended services ranked according to different dimensions; the prior art analyzes the usage of operator services in the home dimension and then calculates more suitable recommended products from multiple dimensions; however, when the number of people using a package is one person, the family package is not applicable, and the special situation of a single person using a package is not considered, resulting in the recommended package not meeting the actual needs of users and affecting the user experience.

[0004] The present invention provides a personalized service system and method based on big data to solve the above technical problems. Summary of the Invention

[0005] The present invention aims to solve at least one of the technical problems existing in the prior art; for this purpose, the present invention proposes a personalized service system and method based on big data, which is used to solve the technical problem that when the number of people using a package is one person, the family package is not applicable, and the special situation of a single person using a package is not considered, resulting in the recommended package not meeting the actual needs of users and affecting the user experience.

[0006] To achieve the above object, a first aspect of the present invention provides a personalized service system based on big data, including: a user analysis module, and a data collection module and an intelligent recommendation module connected thereto;

[0007] Data collection module: used to collect the basic information, real-time behavior data and historical business data of users through several channels;

[0008] User analysis module: used to construct a user portrait according to the basic information of the user; determine the business handled by the user according to the user portrait and real-time behavior data; analyze the personality data of the user according to the historical business data of the user;

[0009] Intelligent recommendation module: used to recommend corresponding business packages for users based on the personality data of the users; monitor the real-time traffic data of the users; recommend value-added services according to the real-time traffic of the users.

[0010] Preferably, the constructing of the user portrait according to the basic information of the user includes:

[0011] Retrieve the basic information of the user; wherein, the basic information includes: name, age, occupation and residential address;

[0012] Convert the residential address in the basic information of the user into longitude and latitude; obtain the residential population table; match the longitude and latitude with the residential population table to obtain the number of family members of the user;

[0013] Integrate the age and occupation in the basic information of the user into a portrait input sequence; call the portrait analysis model; input the portrait input sequence into the portrait analysis model to obtain the corresponding user tags; integrate the user tags with the number of family members of the user into a user portrait; wherein, the portrait analysis model is constructed based on an artificial intelligence model.

[0014] It should be noted that the user tags represent the characteristics of the user. For example, the user tags of user A are "student, young person"; the user tags of user B are "working person, low-consumption group".

[0015] The present invention infers the number of family members based on the residential address in the basic information of the user, analyzes the age and occupation in the basic information using the portrait analysis model to obtain user tags; integrates the user tags and the number of family members into a user portrait; which lays a foundation for subsequent analysis of user packages and is beneficial to improving the accuracy of package recommendation.

[0016] Preferably, the portrait analysis model is constructed based on an artificial intelligence model, including:

[0017] Obtain a standard data set; wherein, the standard data set includes standard output data consistent with the content attributes of the portrait input sequence, and standard output data consistent with the content attributes of the user tags;

[0018] Divide the standard data set into a training set, a validation set, and a test set according to a set ratio; use the training set to train the artificial intelligence model; use the validation set to adjust the internal parameters of the artificial intelligence model; use the test set to test the artificial intelligence model to obtain test metrics;

[0019] Judge whether the test metric is greater than the metric threshold; if yes, mark the trained artificial intelligence model as a portrait analysis model; if not, retrain the artificial intelligence model.

[0020] It should be noted that the test metrics include accuracy, F1 score, recall rate, and stability; the metric threshold and the set ratio are set by expert evaluation; when the artificial intelligence model needs to be retrained, the ratio of the standard data division is re-evaluated, and then the artificial intelligence model is retrained.

[0021] Preferably, the determining the business handled by the user according to the user portrait and real-time behavior data includes:

[0022] Retrieve the real-time behavior data of the user; wherein, the real-time behavior data includes: page stay duration, number of link clicks, and search keywords;

[0023] Extract the user's behavior characteristics from the real-time behavior data, and analyze the correlation between the behavior characteristics and several business types; select the business type with the highest correlation as the business type handled by the user;

[0024] Retrieve the user portrait, and obtain the handling packages in the corresponding business type handled; screen the handling packages according to the user portrait, and integrate the screened several handling packages to obtain the business handled by the user.

[0025] The present invention analyzes the correlation between the user and several business types according to the user's behavior data, analyzes the type of business handled by the user according to the correlation; screens several packages from the business type handled according to the user portrait, and integrates the several packages to obtain the business handled; can accurately analyze the user's needs, is conducive to screening suitable packages according to the user's needs, and improves the user's satisfaction.

[0026] Preferably, the extracting the user's behavior characteristics from the real-time behavior data includes:

[0027] When the real-time behavior data includes search keywords, obtain the key feature library; match the search keywords with the key feature library to obtain the feature information of the search keywords; if the search keywords are not matched with the key feature library, calculate the matching degree between the search keywords and the feature information in the key feature library, and select the feature information with the highest matching degree as the feature information of the search keywords;

[0028] When the real-time behavior data includes the number of link clicks, the links with the number of clicks exceeding the threshold are integrated to obtain the analysis links; the analysis links are matched with the link feature library to obtain the corresponding link feature information;

[0029] When the real-time behavior data includes the page stay duration, the pages with the stay duration exceeding the duration threshold are captured, and the captured pages are converted into extracted text; the N-gram model is used to extract the features in the extracted text to obtain the page feature information;

[0030] The feature information, the link feature information, and the page feature information are integrated to obtain the user's behavior characteristics.

[0031] It should be noted that the user's real-time behavior data may not include all types, and only the data included in the real-time behavior data needs to be analyzed; when the real-time behavior data includes all types, the search keywords, the number of link clicks, and the page stay duration are analyzed in sequence.

[0032] Preferably, analyzing the relevance between the behavior characteristics and several business types includes:

[0033] Retrieving the user's behavior characteristics; using a statistical model to analyze the probabilities P(x) and P(y) of the sub-data and the business type appearing alone in the behavior characteristics, and the probability P(x,y) of the sub-data and the business type appearing simultaneously in the behavior characteristics; calculating the individual relevance between the sub-data in the behavior characteristics and several business types respectively through the formula DXG = P(x,y) / [P(x)×P(y)];

[0034] Mark the individual relevance between the feature information, the link feature information, and the page feature information in the behavior characteristics and several business types as TYi, LYi, and YWi respectively; calculate the relevance between the behavior characteristics and the business type through the formula ZXGi = α×TYi + β×LYi + γ×YWi; where α, β, and γ are weight coefficients greater than 0 respectively; i represents the numbers of several businesses, i = 1, 2,..., n, and n is a positive integer.

[0035] It should be noted that when there is only one sub-data in the behavior characteristics, the corresponding individual relevance is used as the relevance between the behavior characteristics and several business characteristics; when the corresponding sub-data is not included in the behavior characteristics, the individual relevance and the weight coefficient corresponding to the sub-data are marked as 0; the setting of the weight coefficient is set according to the actual situation. When the user's search frequency is relatively high, the corresponding weight coefficient α is set to be larger; when the user's link click frequency is relatively high, the corresponding weight coefficient β is set to be larger; when the user's page browsing frequency is relatively high, the corresponding weight coefficient γ is set to be larger.

[0036] Preferably, analyzing the user's personality data based on the user's historical service data includes:

[0037] Retrieving the user's historical service data; wherein, the historical service data includes: the maximum traffic usage, the total call duration, and the value-added service subscription record;

[0038] Taking the maximum traffic usage in the historical service data as the traffic usage threshold; taking the total call duration in the historical service data as the call duration threshold;

[0039] When there is value-added service cancellation information in the value-added service subscription record, add a label of no value-added service to the user; when there is a subscription to a data package in the value-added service subscription record, calculate the sum of the maximum traffic and the subscribed data package as the new traffic usage threshold;

[0040] When there is other subscription information in the value-added service subscription record, use the corresponding subscription information as the added label for the user; integrate the user's traffic usage threshold, call duration threshold, and added labels into the user's personality data.

[0041] The present invention analyzes the user's personality data based on the user's historical service data, analyzes the user's usage situation according to the user's historical service data, provides a data basis for subsequent recommended services, and is conducive to selecting different packages for different users to meet the user's needs.

[0042] Preferably, recommending a corresponding service package to the user based on the user's personality data includes:

[0043] Retrieving the user's personality data and the services to be handled; using the traffic usage threshold and call duration threshold in the personality data as criteria to determine whether the data of the service package to be handled is greater than the criteria. If so, retain the corresponding service package; if not, eliminate the corresponding service package;

[0044] Determine whether the added label in the personality data is no value-added service. If so, eliminate the service packages containing value-added services; if not, screen out the corresponding service packages according to the subscription information labels added in the personality label;

[0045] Obtain the package prices of the retained service packages, sort the service packages in ascending order of package price to obtain a package recommendation ranking list; select the top n service packages in the package recommendation ranking list as the service packages and recommend the service packages to the corresponding users.

[0046] The present invention screens out corresponding service packages from the business handling based on the user's personality data, sorts the service packages according to the prices of the service packages, and recommends them to the corresponding users; considering the user's personality data and the package price, recommending the service packages is beneficial to making the recommended packages more meet the needs of customers.

[0047] Preferably, the recommending value-added services according to the user's real-time traffic includes:

[0048] Retrieving the user's real-time traffic; obtaining the traffic threshold in the user's service package; comparing the real-time traffic with the traffic threshold; when the real-time traffic reaches the traffic threshold, triggering a pop-up window for recommending value-added services;

[0049] When the real-time traffic does not reach the traffic threshold after a set time period, calculating the difference between the traffic threshold and the real-time traffic to obtain a traffic difference; determining whether the traffic difference is greater than the difference threshold; if yes, triggering a pop-up window for package replacement according to the user's real-time traffic; if no, continuously detecting the user's real-time traffic.

[0050] The present invention monitors the user's real-time traffic. When the user's real-time traffic reaches the traffic threshold, value-added services are recommended to the user; otherwise, the user is recommended to change the package according to the difference between the real-time traffic and the traffic threshold; it can start from the user's actual usage situation, reduce unnecessary expenses of the user, and is beneficial to increasing the user's satisfaction and loyalty.

[0051] The second aspect of the present invention provides a personalized service method based on big data, including:

[0052] Step S1: Obtaining the user's basic information, real-time behavior data, and historical business data through several channels;

[0053] Step S2: Constructing a user portrait according to the user's basic information; determining the user's business to be handled according to the user portrait and real-time behavior data;

[0054] Step S3: Analyzing the user's personality data according to the user's historical business data; recommending corresponding service packages to the user based on the user's personality data;

[0055] Step S4: Monitoring the user's real-time traffic data; recommending value-added services according to the user's real-time traffic.

[0056] Compared with the prior art, the beneficial effects of the present invention are:

[0057] 1. The present invention infers the number of family members based on the residential address in the user's basic information, analyzes the age and occupation in the basic information using a portrait analysis model to obtain user tags; integrates the user tags and the number of family members into a user portrait, laying a foundation for subsequent package analysis of the user, which is beneficial to improving the accuracy of package recommendation; analyzes the correlation between the user and several service types according to the user's behavior data, and analyzes the types of services handled by the user according to the correlation; screens several packages from the types of services handled according to the user portrait, and integrates the several packages to obtain the services to be handled; can accurately analyze the user's needs, is beneficial to screening suitable packages according to the user's needs, and improves the user's satisfaction.

[0058] 2. The present invention analyzes the user's personality data based on the user's historical service data, analyzes the user's usage situation according to the user's historical service data, providing a data basis for subsequent service recommendation, which is beneficial to selecting different packages for different users to meet the user's needs; screens the corresponding handled packages from the services to be handled according to the user's personality data, and sorts and recommends the handled packages to the corresponding users according to the prices of the handled packages; recommends the handled packages considering the user's personality data and the package prices, which is beneficial to making the recommended packages more meet the customer's needs; monitors the user's real-time traffic, and when the user's real-time traffic reaches the traffic threshold, recommends value-added services to the user; otherwise, recommends the user to change the package according to the difference between the real-time traffic and the traffic threshold; can start from the user's actual usage situation, reduce the user's unnecessary expenses, and is beneficial to increasing the user's satisfaction and loyalty. BRIEF DESCRIPTION OF THE DRAWINGS

[0059] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for use in the description of the embodiments or the prior art. Obviously, the following drawings are only some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0060] Figure 1 It is a schematic diagram of the overall steps of the system of the present invention;

[0061] Figure 2 It is a schematic diagram of the steps of constructing the user portrait and dividing the services to be handled of the present invention;

[0062] Figure 3 It is a schematic diagram of the steps of package recommendation and traffic monitoring of the present invention;

[0063] Figure 4 It is a schematic diagram of the specific steps of the method of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0064] The technical solution of the present invention will be clearly and completely described below in conjunction with the embodiments. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative work shall fall within the protection scope of the present invention.

[0065] Please refer to Figure 1 , an embodiment of the first aspect of the present invention provides a personalized service system based on big data, including: a user analysis module, and a data collection module and an intelligent recommendation module connected thereto;

[0066] The data collection module: is used to collect the basic information, real-time behavior data and historical business data of users through several channels;

[0067] The user analysis module: is used to construct a user portrait according to the basic information of the user; determine the business handled by the user according to the user portrait and real-time behavior data; analyze the personality data of the user according to the historical business data of the user;

[0068] The intelligent recommendation module: is used to recommend corresponding business packages to users based on the personality data of the users; monitor the real-time traffic data of the users; recommend value-added services according to the real-time traffic of the users.

[0069] Please refer to Figure 2 , obtain the basic information of the user; wherein, the basic information includes: name, age, occupation and residential address; convert the residential address in the basic information of the user into longitude and latitude; obtain the residential population table; match the longitude and latitude with the residential population table to obtain the number of family members of the user; integrate the age and occupation in the basic information of the user into a portrait input sequence; call the portrait analysis model, and input the portrait input sequence into the portrait analysis model to obtain the corresponding user tags; integrate the user tags with the number of family members of the user into a user portrait; wherein, the portrait analysis model is constructed based on an artificial intelligence model;

[0070] Obtain the real-time behavior data of the user; wherein, the real-time behavior data includes: page stay duration, link click times and search keywords; when the real-time behavior data includes search keywords, then obtain the key feature library; match the search keywords with the key feature library to obtain the feature information of the search keywords; if the search keywords do not match the key feature library, then calculate the matching degree between the search keywords and the feature information in the key feature library, and select the one with the highest matching degree as the feature information of the search keywords;

[0071] When the real-time behavior data includes the number of link clicks, the links with the number of clicks exceeding the threshold are integrated to obtain analysis links; the analysis links are matched with the link feature library to obtain the corresponding link feature information; when the real-time behavior data includes the page stay duration, the pages with the stay duration exceeding the duration threshold are captured, and the captured pages are converted into extracted text; the N-gram model is used to extract the features in the extracted text to obtain the page feature information; the feature information, the link feature information, and the page feature information are integrated to obtain the user's behavior characteristics.

[0072] Use a statistical model to analyze the probabilities P(x) and P(y) of the sub-data and business types appearing alone in the behavior characteristics, and the probability P(x,y) of the sub-data and business type appearing simultaneously in the behavior characteristics; calculate the individual correlations between the sub-data in the behavior characteristics and several business types respectively through the formula DXG = P(x,y) / [P(x)×P(y)]; mark the individual correlations between the feature information, link feature information, and page feature information in the behavior characteristics and several business types as TYi, LYi, and YWi respectively; calculate the correlation between the behavior characteristics and the business type through the formula ZXGi = α×TYi + β×LYi + γ×YWi; where α, β, and γ are weight coefficients greater than 0 respectively; i represents the numbers of several services, i = 1, 2, …, n, and n is a positive integer; select the business type with the largest correlation as the user's handled business type; retrieve the user portrait and obtain the handling packages in the corresponding handled business type; screen the handling packages according to the user portrait, and integrate the selected several handling packages to obtain the user's handled business.

[0073] It should be noted that when the probabilities of the sub-data and business type appearing alone in the behavior characteristics are fixed values; the greater the probability of their simultaneous appearance, the stronger the corresponding individual correlation; the setting of the weight coefficients is based on the actual situation. When the user's search frequency is relatively high, the corresponding weight coefficient α is set to be larger; when the user's link click frequency is relatively high, the corresponding weight coefficient β is set to be larger; when the user's page browsing frequency is relatively high, the corresponding weight coefficient γ is set to be larger; this can make the setting of the weight coefficients conform to the actual situation of the user and make the calculated results more accurate.

[0074] It is worth noting that the portrait analysis model is constructed based on an artificial intelligence model and includes:

[0075] Obtain a standard data set; where the standard data set includes standard output data consistent with the content attributes of the portrait input sequence and standard output data consistent with the content attributes of the user tags.

[0076] Divide the standard data set into a training set, a validation set, and a test set according to a set ratio; use the training set to train the artificial intelligence model; use the validation set to adjust the internal parameters of the artificial intelligence model; use the test set to test the artificial intelligence model to obtain test metrics;

[0077] Judge whether the test metrics are greater than the metric threshold; if so, mark the trained artificial intelligence model as a portrait analysis model; if not, retrain the artificial intelligence model.

[0078] It should be noted that the test metrics include accuracy, F1 score, recall rate, and stability; the metric threshold and the set ratio are set by expert evaluation; when the artificial intelligence model needs to be retrained, the ratio of the standard data division is re-evaluated, and then the artificial intelligence model is trained.

[0079] It should be explained that the family members of the user are analyzed according to the user's basic information; the portrait analysis model is used to analyze the user's age and occupation to obtain user tags; the user's family members and user tags are integrated into a user portrait; the business that the user wants to handle is analyzed according to the user's real-time behavior data, and the business handling is preliminarily screened according to the user portrait; it can analyze the actual needs of the user, laying a foundation for subsequent package recommendations.

[0080] Please refer to Figure 3 , to obtain the user's historical service data; among them, the historical service data includes: the maximum traffic usage, the total call duration, and the value-added service subscription record; use the maximum traffic usage in the historical service data as the traffic usage threshold; use the total call duration in the historical service data as the call duration threshold; when there is value-added service cancellation information in the value-added service subscription record, add a no-value-added-service label to the user; when there is a traffic package subscription in the value-added service subscription record, calculate the sum of the maximum traffic and the subscribed traffic package as the new traffic usage threshold; when there is other subscription information in the value-added service subscription record, use the corresponding subscription information as the added label for the user; integrate the user's traffic usage threshold, call duration threshold, and added labels into the user's personalized data;

[0081] Taking the traffic usage threshold and call duration threshold in the personalized data as criteria, determine whether the data of the handled package in the handled service is greater than the criteria. If so, retain the corresponding handled package; if not, eliminate the corresponding handled package. Determine whether the added label in the personalized data is a non-value-added service. If so, eliminate the handled packages containing value-added services; if not, screen out the corresponding handled packages according to the subscribed information label added in the personalized label. Obtain the package prices of the retained handled packages, sort the handled packages in ascending order of package price to obtain a package recommendation ranking list. Select the top n handled packages in the package recommendation ranking list as service packages and recommend the service packages to the corresponding users.

[0082] Monitor the user's real-time traffic data; obtain the traffic threshold in the user's service package. When the real-time traffic reaches the traffic threshold, trigger a pop-up window to recommend value-added services. When the real-time traffic does not reach the traffic threshold after a set time period, calculate the difference between the traffic threshold and the real-time traffic to obtain a traffic difference. Determine whether the traffic difference is greater than the difference threshold. If so, trigger a package replacement pop-up window according to the user's real-time traffic; if not, continuously detect the user's real-time traffic.

[0083] It should be noted that the personalized data of the user is analyzed according to the user's historical service data; the handled packages are screened according to the user's personalized data, and finally the handled packages are recommended to the user according to the package price; and according to the user's traffic usage situation, value-added services are recommended or the user is recommended to change the package, which can increase the user's satisfaction and loyalty.

[0084] Please refer to Figure 4 , the second aspect embodiment of the present invention provides a personalized service method based on big data, including:

[0085] Step S1: Obtain the basic information, real-time behavior data, and historical service data of the user through several channels;

[0086] Step S2: Construct a user portrait according to the user's basic information; determine the user's handled service according to the user portrait and real-time behavior data;

[0087] Step S3: Analyze the user's personalized data according to the user's historical service data; recommend corresponding service packages to the user based on the user's personalized data;

[0088] Step S4: Monitor the user's real-time traffic data; recommend value-added services according to the user's real-time traffic.

[0089] Some of the data in the above formula is calculated by removing the dimension and taking its numerical value. The formula is obtained by software simulation of a large amount of collected data to get a formula closest to the actual situation; the preset parameters and preset thresholds in the formula are set by those skilled in the art according to the actual situation or obtained through simulation of a large amount of data.

[0090] The working principle of the present invention: The present invention obtains the basic information, real-time behavior data and historical business data of users through several channels; constructs a user portrait according to the basic information of the users; determines the business to be handled by the users according to the user portrait and real-time behavior data; analyzes the personality data of the users according to the historical business data of the users; recommends corresponding business packages for the users based on the personality data of the users; monitors the real-time traffic data of the users; recommends value-added services according to the real-time traffic of the users.

[0091] The above embodiments are only used to illustrate the technical method of the present invention and not to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those of ordinary skill in the art should understand that the technical method of the present invention can be modified or equivalently replaced without departing from the spirit and scope of the technical method of the present invention.

Claims

1. A personalized service system based on big data, characterized in that: include: User analysis module, and the data collection module and intelligent recommendation module connected to it; Data collection module: used to collect basic information, real-time behavior data and historical business data of users through several channels; User analysis module: used to build user portraits based on the user's basic information; Determine the user's business based on user portraits and real-time behavior data; Analyze the user's personality data based on the user's historical business data; Intelligent recommendation module: used to recommend corresponding service packages to users based on their personal data; Monitor users' real-time traffic data; Recommend value-added services based on users' real-time traffic.

2. According to claim 1, a personalized service system based on big data is characterized in that: The step of constructing a user portrait based on the user's basic information includes: Retrieve the user's basic information, including name, age, occupation and residential address; Convert the residential address in the user's basic information into longitude and latitude; obtain the residential population table; match the longitude and latitude with the residential population table to obtain the number of family members of the user; Integrate the age and occupation in the user's basic information into a portrait input sequence; call the portrait analysis model; input the portrait input sequence into the portrait analysis model to obtain the corresponding user label; integrate the user label and the number of family members of the user into a user portrait; wherein, the portrait analysis model is built based on an artificial intelligence model.

3. The personalized service system based on big data according to claim 2, characterized in that: The portrait analysis model is constructed based on an artificial intelligence model and includes: Acquire a standard data set; wherein the standard data set includes standard output data consistent with the content attributes of the portrait input sequence and standard output data consistent with the content attributes of the user tag; Divide the standard data set into training set, validation set and test set according to the set ratio; use the training set to train the artificial intelligence model; use the validation set to adjust the internal parameters of the artificial intelligence model; use the test set to test the artificial intelligence model and obtain the test indicators; Determine whether the test index is greater than the index threshold; if yes, mark the trained artificial intelligence model as a portrait analysis model; if no, retrain the artificial intelligence model.

4. The personalized service system based on big data according to claim 1, characterized in that: Determining the user's business according to the user portrait and real-time behavior data includes: Retrieve real-time user behavior data, including page dwell time, link clicks, and search keywords; Extracting the user's behavior characteristics from the real-time behavior data, analyzing the correlation between the behavior characteristics and several business types; selecting the business type with the greatest correlation as the user's business type; Retrieve the user portrait and obtain the service packages of the corresponding service type; filter the service packages according to the user portrait, integrate the filtered service packages and obtain the user's service.

5. The personalized service system based on big data according to claim 4, characterized in that: The extracting of user behavior characteristics from real-time behavior data includes: When the real-time behavior data includes a search keyword, a key feature library is obtained; the search keyword is matched with the key feature library to obtain feature information of the search keyword; if the search keyword does not match with the key feature library, the degree of matching between the search keyword and the feature information in the key feature library is calculated, and the feature information with the highest matching degree is selected as the feature information of the search keyword; When the real-time behavior data includes the number of link clicks, the links whose clicks exceed the number threshold are integrated to obtain analysis links; the analysis links are matched with the link feature library to obtain corresponding link feature information; When the real-time behavior data includes the page dwell time, the page whose dwell time exceeds the time threshold is screenshotted, and the screenshot page is converted into extracted text; the features in the extracted text are extracted using the N-gram model to obtain page feature information; Integrate feature information, link feature information and page feature information to obtain user behavior characteristics.

6. The personalized service system based on big data according to claim 4, characterized in that: The analysis behavior characteristics and the correlation with several business types include: Retrieve the user's behavioral characteristics; use statistical models to analyze the probability P(x) and P(y) of the sub-data and business type appearing separately in the behavioral characteristics, as well as the probability P(x, y) of the sub-data and business type appearing simultaneously in the behavioral characteristics; calculate the individual correlations between the sub-data in the behavioral characteristics and several business types respectively through the formula DXG=P(x, y) / [P(x)×P(y)]; The individual correlations between the characteristic information, link characteristic information and page characteristic information in the behavior characteristics and several business types are marked as TYi, LYi and YWi respectively; the correlation between the behavior characteristics and the business type is calculated by the formula ZXGi=α×TYi+β×LYi+γ×YWi; wherein α, β and γ are weight coefficients greater than 0 respectively; i represents the number of several businesses, i=1,2,…,n, and n is a positive integer.

7. The personalized service system based on big data according to claim 1, characterized in that: Analyzing the user's personality data based on the user's historical business data includes: Retrieve the user's historical service data; the historical service data includes: maximum traffic usage, total call duration, and value-added service subscription records; The maximum traffic usage value in the historical service data is used as the traffic usage threshold; the total call duration in the historical service data is used as the call duration threshold; When there is VAS unsubscription information in the VAS subscription record, a tag of no VAS is added to the user; when there is a traffic package subscription in the VAS subscription record, the sum of the calculated maximum traffic value and the subscribed traffic package is used as the new traffic usage threshold; When other subscription information exists in the value-added service subscription record, the corresponding subscription information is used as the user's added tag; the user's traffic usage threshold, call duration threshold and added tag are integrated into the user's personalized data.

8. The personalized service system based on big data according to claim 1, characterized in that: The recommending corresponding service packages to the user based on the user's personality data includes: Retrieve the user's personal data and business processing; use the traffic usage threshold and call duration threshold in the personal data as the standard to determine whether the data of the processing package in the processing business is greater than the standard. If yes, retain the corresponding processing package; if not, remove the corresponding processing package; Determine whether the added tag in the personal data is no value-added service; if yes, remove the service package containing the value-added service; if no, select the corresponding service package according to the subscription information tag added in the personal tag; Obtain the package prices of the reserved processing packages, sort the processing packages from low to high according to the package prices, and obtain a package recommendation sorting table; select the first n processing packages in the package recommendation sorting table as business packages, and recommend the business packages to corresponding users.

9. The personalized service system based on big data according to claim 1, characterized in that: The recommending of value-added services according to the user's real-time traffic includes: Retrieve the user's real-time traffic; obtain the traffic threshold in the user's service package; compare the real-time traffic with the traffic threshold; when the real-time traffic reaches the traffic threshold, trigger a pop-up window to recommend value-added services; When the real-time traffic does not reach the traffic threshold after the set time period, the difference between the traffic threshold and the real-time traffic is calculated to obtain the traffic difference; determine whether the traffic difference is greater than the difference threshold; if so, trigger a package change pop-up window based on the user's real-time traffic; if not, continue to detect the user's real-time traffic.

10. A personalized service method based on big data, applied to a big data personalized service system according to any one of claims 1 to 9, characterized in that: include: Step S1: basic information, real-time behavior data and historical business data of users from several channels; Step S2: construct a user profile based on the user's basic information; Determine the user's business based on user portraits and real-time behavior data; Step S3: analyzing the user's personality data according to the user's historical service data; and recommending a corresponding service package to the user based on the user's personality data; Step S4: monitoring the user's real-time traffic data; Recommend value-added services based on users' real-time traffic.

Images