Friend-making data analysis and recommendation method and system based on AI

Abstract

The invention discloses an AI-based friend-making data analysis and recommendation method and system, and relates to the field of data analysis, and the main scheme is as follows: calculating a user friend-making preference index X according to a character behavior consistency index XA, a life financial deviation index XB, a user friend closeness XC and a tourism characteristic index XD; calculating friend-making preference similarity Yh according to the friend-making preference index X of the user and friend-making preference indexes Yh of other users, and judging friend-making preference similarity levels according to the friend-making preference similarity Yh and a friend-making preference similarity threshold set, thereby solving the problems that data types for evaluating the friend-making preferences of the user are single and lack of diversity, meanwhile, the relation among various types of data is ignored, and the evaluation accuracy of the friend-making preferences of the user is improved. And therefore, the accuracy of friend matching is reduced.

Description

Technical Field

[0001] The present invention relates to the field of data analysis technology, and specifically to an AI-based dating data analysis and recommendation method and system. Background Art

[0002] In the field of social dating, through the analysis of dating data, we can build accurate portraits, make recommendations of more suitable dating partners, improve the matching success rate, and make the dating process richer and smoother.

[0003] The working principle of a traditional AI-based dating data analysis and recommendation method and system is: collect users' basic personal information such as age, gender, occupation, etc. from multiple channels, analyze and calculate the collected information, obtain the user's dating preferences, and make friend recommendations based on the user's dating preferences.

[0004] The existing technology still has the following shortcomings: the data type for evaluating users' friendship preferences is single and lacks diversity, and the connection between various types of data is ignored, which leads to a decrease in the accuracy of friend matching. Summary of the invention

[0005] 1. Technical issues to be solved

[0006] In view of the shortcomings of the prior art, the present invention provides an AI-based dating data analysis and recommendation method and system, which calculates the user dating preference index X according to the personality behavior consistency index XA, the life financial deviation index XB, the user friend closeness XC and the travel feature index XD, and calculates the user dating preference index X according to the user dating preference index X and other user dating preference indexes X h Calculate the similarity Y of friendship preferences h , according to the similarity Y of friendship preference h The similarity level of friendship preferences is judged by the friendship preference similarity threshold set, which solves the problem that the data type for evaluating user friendship preferences is single and lacks diversity, while ignoring the connection between various types of data, thereby reducing the accuracy of friend matching.

[0007] (II) Technical solution

[0008] To achieve the above objectives, the present invention is implemented through the following technical solutions: an AI-based dating data analysis and recommendation method, comprising the following steps:

[0009] Step 1: Collect the user's personality type data, social behavior data, life data, financial data and travel data;

[0010] Step 2: Calculate the personality behavior consistency index XA based on the personality type data and the social behavior data; calculate the life balance index BA based on the life data, calculate the consumption adaptation life rhythm index BB based on the life data and the financial data, and calculate the life finance deviation index XB based on the life balance index BA and the consumption adaptation life rhythm index BB; calculate the user's friend closeness XC based on the social behavior data; calculate the travel characteristic index XD based on the social behavior data and the travel data; calculate the user's friendship preference index X based on the personality behavior consistency index XA, the life finance deviation index XB, the user's friend closeness XC and the travel characteristic index XD;

[0011] Step 3: Calculate the friendship preference index X of other users according to the method in steps 1-2 h , based on user friendship preference index X and other user friendship preference index X h Calculate the similarity Y of friendship preferences h ;

[0012] Step 4: Preset a threshold set of similarity levels of friendship preferences; h The friendship preference similarity level is determined by the friendship preference similarity threshold set, and the friend recommendation priority is determined according to the friendship preference similarity level.

[0013] In the preferred embodiment of the above-mentioned AI-based friendship data analysis and recommendation method, the method for calculating the personality behavior consistency index XA is:

[0014] Personality type data includes personality expected behavior value A 1 ;

[0015] Social behavior data includes the number of times a user actively initiates a chat per unit time. 2 、Total number of chats A 3 、Number of offline activities participated inA 4 And the total number of activities that the platform pushes to users that can participate A 5 ;

[0016] The personality behavior consistency index XA is calculated based on personality type data and social behavior data, and the formula is:

[0017]

[0018] In the preferred embodiment of the above-mentioned AI-based dating data analysis and recommendation method, the method for calculating the life balance index BA and the consumption adaptation life rhythm index BB is:

[0019] Life data including standard deviation of wake-up time Standard deviation of sleep time Standard deviation of work and study time ratio Standard deviation of leisure and entertainment time Standard deviation of time spent on family affairs Standard deviation of social time Frequency of work and study activities score B 7 , Leisure and entertainment activity frequency score B 8 , Family activity frequency score B 9 and social activity frequency score B 10 ;

[0020] Financial data includes consumption amount D a ;

[0021] Based on the standard deviation of wake-up time Standard deviation of sleep time Standard deviation of work and study time ratio Standard deviation of leisure and entertainment time Standard deviation of time spent on family affairs and social time percentage standard deviation The calculation formula for life balance index BA is:

[0022]

[0023] Score B based on frequency of work and study activities 7 , Leisure and entertainment activity frequency score B 8 , Family activity frequency score B 9 , Social activity frequency score B 10 and consumption amount D a The formula for calculating the consumption adaptation life rhythm index BB is as follows:

[0024]

[0025] Among them, D a is the amount of money spent by the user in the ath consumption category, a is the serial number corresponding to different consumption categories, and its value is [1, b]; b is the total number of consumption categories, and its value is a positive integer.

[0026] In the preferred embodiment of the above-mentioned AI-based dating data analysis and recommendation method, the method for calculating the life financial deviation index XB is:

[0027] The life financial deviation index XB is calculated based on the life balance index BA and the consumption adaptation life rhythm index BB, and the formula is as follows:

[0028]

[0029] Among them, BA max The maximum value of all user life balance indicators; BA minThe minimum value of all user life balance indicators; BB max The maximum value of the life rhythm index for all users; BB min The minimum value of the life rhythm index for all users' consumption adaptation.

[0030] In the preferred solution of the above-mentioned AI-based dating data analysis and recommendation method: the method for calculating the user's friend closeness XC is:

[0031] Social behavior data also includes the number of friends EA d , Number of users EB d And the number of user friend nodes EC d ;

[0032] EA based on the number of friends d , Number of users EB d And the number of user friend nodes EC d The formula for calculating the user's friend closeness XC is:

[0033]

[0034] Among them, EA d is the number of friends of the user on the dth social software, d is the number of different social software, and its value is [1, i]; i is the total number of social software used by the user, and its value is a positive integer; EB d is the number of users using the dth social software, which is a positive integer; EC d is the number of user friend nodes in the dth social software.

[0035] In the preferred embodiment of the above-mentioned AI-based dating data analysis and recommendation method, the method for calculating the tourism characteristic index XD is:

[0036] Social behavior data also includes the number of travel updates published ED;

[0037] Travel data includes travel expenditure amount FA, total travel time FB and travel play time FC;

[0038] The tourism characteristic index XD is calculated based on the number of travel dynamics releases ED and travel data, and the formula is as follows:

[0039]

[0040] Among them, α 1 is the weight coefficient of tourism consumption efficiency index, ranging from 0.3 to 0.7; α 2 is the weight coefficient of the tourism activity index, ranging from 0.3 to 0.7; and α 1 +α 2 =1.

[0041] In the preferred embodiment of the above-mentioned AI-based dating data analysis and recommendation method, the method for calculating the user dating preference index X is:

[0042] The user friendship preference index X is calculated based on the personality behavior consistency index XA, the life financial deviation index XB, the user friend closeness XD and the travel characteristic index XD. The formula is:

[0043]

[0044] Among them, β 1 is the weight coefficient of the personality-behavior consistency index XA, ranging from 0.2 to 0.4; β 2 is the weight coefficient of the life finance deviation index XB, ranging from 0.2 to 0.3; β 3 is the weight coefficient of the user's friend closeness XC, ranging from 0.2 to 0.3; β 4 is the weight coefficient of tourism characteristic index XD, ranging from 0.3 to 0.4; and β 1 +β 2 +β 3 +β 4 =1.

[0045] In the preferred embodiment of the above-mentioned AI-based dating data analysis and recommendation method: calculating the dating preference similarity Y h The method is:

[0046] Based on user friendship preference index X and other user friendship preference index X h Calculate the similarity Y of friendship preferences h , the formula based on is:

[0047]

[0048] Among them, X h is the user friendship preference index of the hth user on the platform, h is the serial number corresponding to different users, and its value is a positive integer; Y h is the similarity of the user's friendship preference with the hth user on the platform.

[0049] In the preferred embodiment of the above-mentioned AI-based dating data analysis and recommendation method, the method for determining the similarity level of dating preferences is:

[0050] The friendship preference similarity level threshold set includes a high similarity threshold YG and a low similarity threshold YD;

[0051] According to the similarity of friendship preference Y h The formula for judging the similarity level of friendship preference based on the threshold set of friendship preference similarity level is as follows:

[0052]

[0053] The priority of friend recommendation is determined based on the similarity level of friendship preferences, specifically: high similarity level > medium similarity level > low similarity level.

[0054] The present invention also discloses an AI-based dating data analysis and recommendation system, comprising:

[0055] Data collection module, which can collect user's personality type data, social behavior data, life data, financial data and travel data;

[0056] The index calculation module can calculate the personality behavior consistency index XA based on personality type data and social behavior data; calculate the life balance index BA based on life data, calculate the consumption adaptation life rhythm index BB based on life data and financial data, and calculate the life finance deviation index XB based on the life balance index BA and the consumption adaptation life rhythm index BB; calculate the user's friend closeness XC based on social behavior data; calculate the travel characteristic index XD based on social behavior data and travel data; calculate the user's friendship preference index X based on the personality behavior consistency index XA, the life finance deviation index XB, the user's friend closeness XC and the travel characteristic index XD;

[0057] The friendship preference similarity calculation module can calculate other users' friendship preference index X according to the above method. h , based on user friendship preference index X and other user friendship preference index X h Calculate the similarity Y of friendship preferences h ;

[0058] The recommendation module can preset a threshold set of similarity levels of friendship preferences; according to the friendship preference similarity Y h The friendship preference similarity level is determined by the friendship preference similarity threshold set, and the friend recommendation priority is determined according to the friendship preference similarity level.

[0059] (III) Beneficial effects

[0060] The present invention provides an AI-based dating data analysis and recommendation method and system, which has the following beneficial effects:

[0061] (1) By collecting users’ personality type data, social behavior data, life data, financial data, and travel data, it can provide multi-dimensional data for evaluating users’ friendship preferences and improve the accuracy of the evaluation.

[0062] (2) The personality behavior consistency index XA calculated based on personality type data and social behavior data can accurately grasp the behavioral pattern tendencies of users based on their personality traits in interpersonal communication. The life and financial deviation index XB calculated based on life data and financial data can present the coordination status of users' financial health and consumption and life rhythm in daily life. The user's friend closeness XC calculated based on social behavior data can clearly understand the depth and closeness of the user's relationship in the existing social network. The travel characteristic index XD calculated based on social behavior data and travel data is helpful to discover the user's social needs in travel scenarios and improve the richness and fun of making friends. The user's friendship preference index X calculated based on the personality behavior consistency index XA, the life and financial deviation index XB, the user's friend closeness XC and the travel characteristic index XD can provide a more comprehensive basis for the friendship recommendation system, reduce invalid social interactions caused by blind recommendations, and improve the success rate of friendship matching.

[0063] (3) Based on user friendship preference index X and other user friendship preference index X h Calculate the similarity Y of friendship preferences h , which can accurately match users with similar interests and values, reducing the number of friendship failures caused by lack of common topics or differences in ideas.

[0064] (4) Based on the similarity Y of friendship preferences h Judging the similarity level of dating preferences by the dating preference similarity threshold set and determining the friend recommendation priority according to the dating preference similarity level can more reasonably allocate platform resources and help the platform better understand user needs and behavior patterns, thereby optimizing the platform's operating strategy. BRIEF DESCRIPTION OF THE DRAWINGS

[0065] Figure 1 A schematic diagram of the working steps of an AI-based dating data analysis and recommendation method of the present invention;

[0066] Figure 2 A schematic diagram of the steps of calculating the life finance deviation index XB in an AI-based friendship data analysis and recommendation method of the present invention;

[0067] Figure 3 The present invention is a structural diagram of an AI-based friendship data analysis and recommendation system. DETAILED DESCRIPTION

[0068] 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.

[0069] See also Figure 1-2 The present invention provides an AI-based dating data analysis and recommendation method, comprising the following steps:

[0070] Step 1: Collect the user's personality type data, social behavior data, life data, financial data and travel data.

[0071] Combined with the content of step 1:

[0072] By collecting users' personality type data, social behavior data, life data, financial data and travel data, we can provide multi-dimensional data for evaluating users' friendship preferences and improve the accuracy of the evaluation.

[0073] Step 2: Calculate the personality behavior consistency index XA based on the personality type data and social behavior data; calculate the life balance index BA based on the life data, calculate the consumption adaptation life rhythm index BB based on the life data and financial data, and calculate the life finance deviation index XB based on the life balance index BA and the consumption adaptation life rhythm index BB; calculate the user friend closeness XC based on the social behavior data; calculate the travel characteristic index XD based on the social behavior data and travel data; calculate the user friendship preference index X based on the personality behavior consistency index XA, the life finance deviation index XB, the user friend closeness XC and the travel characteristic index XD.

[0074] Step 2 includes the following steps:

[0075] Step 201: The method for calculating the personality behavior consistency index XA is:

[0076] Personality type data includes personality expected behavior value A 1 .

[0077] It should be noted that the expected behavior value of personality is A 1It is a quantitative indicator used to measure the degree of behavioral tendency that an individual may have in a specific situation based on his or her personality traits. The method of obtaining it is: the user fills in the Likert scale. The Likert scale is a psychological measurement scale that plays an important role in studying an individual's attitudes, values ​​and behavioral intentions. It consists of a series of statements. The user chooses on an ordered scale based on the degree of agreement with each statement. This scale includes five options, namely strongly disagree, disagree, neutral, agree and strongly agree. The score corresponding to strongly disagree is 1 point, the score corresponding to disagree is 2 points, the score corresponding to neutral is 3 points, the score corresponding to agree is 4 points, and the score corresponding to strongly agree is 5 points. The total score is calculated based on the user's selection results and divided by the number of choices as the personality expected behavior value A. 1 .

[0078] Social behavior data includes the number of times a user actively initiates a chat per unit time. 2 、Total number of chats A 3 、Number of offline activities participated inA 4 And the total number of activities that the platform pushes to users that can participate A 5 .

[0079] It should be noted that the number of times a user actively initiates a chat is A 2 It refers to the number of chat conversations initiated by users within a unit time, reflecting the user's enthusiasm and willingness to actively socialize on the social platform. The acquisition method is as follows: the server of the social platform can record the timestamp of each user sending a chat request, and count the number of chat request records within a set unit time and calculate their sum as the number of users actively initiating chats A. 2 . Total number of chats A 3 It refers to the total number of chats initiated by the user and the number of chats responded to after receiving chats initiated by other users. It reflects the overall activeness of users in participating in chat interactions on the platform. The method of obtaining it is: by querying the chat record database related to the target user, calculating the total number of records in a unit time, and taking it as the total number of chats A. 3 Number of offline activities participated in A 4 It refers to the number of times users actually participate in offline social activities organized by social platforms or spontaneously organized by platform users. It can reflect the user's enthusiasm for extending online social relationships to offline. It is an important indicator for measuring the depth of user participation in social platform activities. By collecting the check-in data of offline activities, the number of user check-ins per unit time is counted and the sum is calculated as the number of offline activities participated in. 4 The platform pushes the total number of activities that users can participate in A 5It refers to the number of all activities that the social platform pushes to users in a unit of time. The acquisition method is as follows: when the platform pushes activity information, it will record the relevant data pushed in the database, including push time, push content, push object, etc. The number of activity information pushed to the target user in a unit of time is counted and the sum is calculated as the total number of activities pushed to users by the platform A 5 The unit time is determined according to user needs and can be one day, one week, one month, one quarter or one year, etc.

[0080] The personality behavior consistency index XA is calculated based on personality type data and social behavior data, and the formula is:

[0081]

[0082] It should be noted that the operating principle of this formula is: It is the actual behavior indicator of users. By calculating the absolute difference between the expected behavior and the actual behavior, the degree of deviation between the two can be reflected. It is used to normalize the difference to obtain the personality behavior consistency index XA. The smaller the deviation between the expected personality behavior and the actual behavior, the higher the personality behavior consistency index XA.

[0083] Step 202: The method for calculating the life balance index BA and the consumption adaptation life rhythm index BB is:

[0084] Life data including standard deviation of wake-up time Standard deviation of sleep time Standard deviation of work and study time ratio Standard deviation of leisure and entertainment time Standard deviation of time spent on family affairs Standard deviation of social time Frequency of work and study activities score B 7 , Leisure and entertainment activity frequency score B 8 , Family activity frequency score B 9 and social activity frequency score B 10 .

[0085] It should be noted that the standard deviation of the wake-up time It reflects the fluctuation of the user's wake-up time within a unit time. The method of obtaining it is: through the alarm record in the user's mobile phone system, the time of checking the mobile phone for the first time every day, or the sleep data recorded by wearable devices such as smart bracelets and smart watches, the user's wake-up time data within a unit time is obtained and its standard deviation is calculated as the wake-up time standard deviation

[0086] Standard deviation of sleep time It is used to measure the degree of change in the user's sleep time, reflecting the stability of the user's sleep pattern. The method of obtaining it is: obtain the user's sleep time data per unit time through the last screen shutdown time of the user's mobile device or the sleep data recorded by wearable devices such as smart bracelets and smart watches, and calculate its standard deviation as the sleep time standard deviation

[0087] Standard deviation of work and study time ratio It reflects the fluctuation of the proportion of users' work and study time. The proportion of work and study time refers to the proportion of time users spend on work and study in a day. The method of obtaining it is as follows: collect the time users spend on office software and online learning platforms every day, divide the time by 24, as the daily work and study time proportion, collect the data of work and study time proportion per unit time and calculate its standard deviation, as the standard deviation of work and study time proportion

[0088] Standard deviation of leisure and entertainment time It reflects the fluctuation of the proportion of users' leisure and entertainment time. The proportion of leisure and entertainment time refers to the proportion of users' leisure and entertainment activities in a day. Leisure and entertainment activities include watching movies, exercising, playing games, etc. The acquisition method is: collect the time users spend using leisure and entertainment software such as game applications, video playback applications, sports and fitness applications every day, divide the time by 24, as the daily leisure and entertainment time proportion, collect the leisure and entertainment time proportion data per unit time and calculate its standard deviation, as the standard deviation of the leisure and entertainment time proportion

[0089] Standard deviation of time spent on family affairs It reflects the fluctuation of the proportion of time spent on family affairs by users. The proportion of time spent on family affairs refers to the proportion of time spent by users on handling family affairs in a day. Family affairs include housework, taking care of family members, etc. The method of obtaining the time spent on family affairs by users through family affairs management applications such as Jiajibang and Zhaiguanjia, or through the family activity time manually recorded by users, and then dividing the time by 24 as the daily proportion of time spent on family affairs. The proportion of time spent on family affairs per unit time is collected and its standard deviation is calculated as the standard deviation of the proportion of time spent on family affairs

[0090] Standard deviation of social time It reflects the fluctuation of the proportion of users' social time. The proportion of social time refers to the proportion of time users spend on socializing in a day. The acquisition method is as follows: WeChat, QQ, WhatsApp and other chat applications will record the sending and receiving time of chat messages. The time users spend on chat applications every day is counted through the background data interface of these applications or by using third-party data collection tools. The time is divided by 24 as the social time proportion. The social time proportion data per unit time is collected and its standard deviation is calculated as the standard deviation of the social time proportion.

[0091] The standard deviation is calculated as follows: first calculate the average value of the statistical data, then calculate the sum of the squares of the difference between each data point and the average value, and finally divide this sum by the total number of data points, and then take the square root of the result to obtain the standard deviation.

[0092] Frequency of work and study activities score B 7 It is used to measure the frequency of users' participation in work and learning activities. The method of obtaining the frequency of work and learning activities is as follows: collect and calculate the total time users spend using office software and online learning platforms per unit time, divide it by the unit time, and get the frequency of work and learning activities. When the frequency of work and learning activities is less than 1 hour per day, the frequency score of work and learning activities is B. 7 1 point, when the frequency of work and study activities is greater than 1 hour per day and less than 3 hours per day, the frequency score of work and study activities is B 7 3 points, when the frequency of work and study activities is greater than 3 hours per day, the frequency score of work and study activities is B 7 is 5 points.

[0093] Frequency of leisure and entertainment activities score B 8 It is used to measure the frequency of users' participation in leisure and entertainment activities. The method of obtaining it is: collect and calculate the total time that users use leisure and entertainment software such as game applications, video playback applications, sports and fitness applications per unit time, and divide it by the unit time to get the frequency of leisure and entertainment activities. When the frequency of leisure and entertainment activities is less than 1 hour per day, the frequency score of leisure and entertainment activities is B. 8 1 point, when the frequency of leisure and entertainment activities is greater than 1 hour per day and less than 3 hours per day, the frequency score of leisure and entertainment activities is B 8 3 points, when the frequency of leisure and entertainment activities is greater than 3 hours a day, the frequency score of leisure and entertainment activities is B 8 is 5 points.

[0094] Family activity frequency score B 9It is used to measure the frequency of users' participation in family activities. The method of obtaining it is as follows: design a questionnaire on the frequency of family activities. The questionnaire clearly lists various family activities, including family dinners, family outings, doing housework together, family meetings, parent-child activities, etc. For each family activity, set five different frequency options, and users can choose according to their own circumstances. The five frequency options are: every day, 1 to 2 times a week, 1 to 2 times a month, 1 to 2 times a quarter, and 1 to 2 times a year. The scores corresponding to the five frequency options are: 1 point, 2 points, 3 points, 4 points, and 5 points. Count the users' choices and calculate their total score, which is then divided by the number of questions to be used as the family activity frequency score B. 9 .

[0095] Social activity frequency score B 10 Used to measure the frequency of users' participation in social activities. The acquisition method is as follows: WeChat, QQ, WhatsApp and other chat applications will record the sending and receiving time of chat messages. Through the background data interface of these applications or using third-party data collection tools to count the time users spend on chat applications every day, calculate the average time per unit time as the frequency of social activities. When the frequency of social activities is less than 1 hour per day, the social activity frequency score is B. 10 When the frequency of social activities is greater than 1 hour per day and less than 3 hours per day, the frequency score of social activities is B. 10 3 points, when the frequency of social activities is greater than 3 hours a day, the frequency score of social activities is B 10 is 5 points.

[0096] Financial data includes consumption amount D a .

[0097] It should be noted that the consumption amount D a It refers to the amount of consumption spent by users in different consumption categories. It is determined by checking the merchant name, transaction description and other information from the bank bill and third-party payment platform records, and classifying the consumption amount into different consumption categories. The consumption categories include: daily necessities, clothing and accessories, housing, transportation, education and learning, medical care, communications and social networking, financial services, etc.

[0098] Based on the standard deviation of wake-up time Standard deviation of sleep time Standard deviation of work and study time ratio Standard deviation of leisure and entertainment time Standard deviation of time spent on family affairs and social time percentage standard deviation The calculation formula for life balance index BA is:

[0099]

[0100] It should be noted that this formula measures the stability of user time allocation by calculating the fluctuations in the user's waking up time, sleeping time, and the time spent participating in various activities. The smaller the fluctuation, the more stable the user's time allocation and the higher the life balance index BA, where e is the base of the natural logarithm in the exponential function.

[0101] Score B based on frequency of work and study activities 7 , Leisure and entertainment activity frequency score B 8 , Family activity frequency score B 9 , Social activity frequency score B 10 The consumption adaptation life rhythm index BB is calculated based on the consumption amount Da, and the formula is:

[0102]

[0103] Among them, D a is the amount of money spent by the user in the ath consumption category, a is the serial number corresponding to different consumption categories, and its value is [1, b]; b is the total number of consumption categories, and its value is a positive integer.

[0104] It should be noted that the operating principle of this formula is: Used to measure whether the user's life rhythm is stable. It is used to measure the consumption diversity of users. By comprehensively considering the two indicators, we get the consumption adaptation life rhythm index BB. The higher the consumption adaptation life rhythm index BB is, the more consistent the user's consumption pattern is with the life rhythm.

[0105] Step 203: The method for calculating the life finance deviation index XB is:

[0106] The life financial deviation index XB is calculated based on the life balance index BA and the consumption adaptation life rhythm index BB, and the formula is as follows:

[0107]

[0108] Among them, BA max is the maximum value among all the life balance indicators of the users, which is determined by calculating all the life balance indicators of the users according to the method in step 202 and screening the maximum value, BA min is the minimum value among all the life balance indices of the users, which is determined by calculating all the life balance indices of the users according to the method in step 202 and screening the minimum value, BB max The maximum value of all user consumption adaptation life rhythm indicators is determined by calculating all user consumption adaptation life rhythm indicators according to the method in step 202 and screening the maximum value, BB minThe minimum value among all user consumption adaptation life rhythm indicators is determined by calculating all user consumption adaptation life rhythm indicators according to the method in step 202 and screening the minimum value.

[0109] It should be noted that this formula obtains the life finance deviation index XB after standardizing the life balance index BA and the consumption adaptation life rhythm index BB and calculating their difference. The life finance deviation index XB is used to measure the inconsistency of users in life balance and consumption adaptation life rhythm. The greater the inconsistency, the worse the coordination between life balance and consumption adaptation life rhythm, and the more likely the user's lifestyle is to mismatch with his finances or consumption habits.

[0110] Step 204: The method for calculating the user's friend closeness XC is:

[0111] Social behavior data also includes the number of friends EA d , Number of users EB d And the number of user friend nodes EC d .

[0112] It should be noted that the number of friends EA d It refers to the number of other users added by a user in different social software, reflecting the size of the user's social network on the social platform. The number of friends is determined through the built-in functions of the social software. d For example, in Facebook, users can see the number of their friends in the "Friends" tab of their personal homepage, and in WeChat, they can check the number of added contacts through the address book. User Number EB d Refers to the total number of registered users of different social software. The official platform will announce it through regular reports or announcements. For example, Weibo will announce the number of active users in its annual financial report. d Refers to the number of mutual friends among the user's friends. The acquisition method is: use the API provided by the official social software to obtain the friend relationship data of the user and the user's friends, use SQL statements to query and calculate, and get the number of user friend nodes EC d Among them, SQL statements are structured query languages, which are standard languages ​​for managing relational databases. They can realize efficient storage, retrieval, update and management of large amounts of data, such as multi-table joint queries, data sorting, grouping statistics, etc.

[0113] EA based on the number of friends d , Number of users EB d And the number of user friend nodes EC d The formula for calculating the user's friend closeness XC is:

[0114]

[0115] Among them, EA d is the number of friends of the user on the dth social software, d is the number of different social software, and its value is [1, i]; i is the total number of social software used by the user, and its value is a positive integer; EB d is the number of users using the dth social software, which is a positive integer; EC d is the number of user friend nodes in the dth social software.

[0116] It should be noted that the operating principle of this formula is: It is used to measure the degree of connection of individual users. It is used to measure the closeness between the friends of a user. By comprehensively considering the influence of two aspects, the user's friend closeness XC is obtained. The user's friend closeness XC reflects the cohesion of the user's social group. The higher the user's friend closeness XC, the higher the cohesion, and the more the user tends to become friends with users related to his social group.

[0117] Step 205: The method for calculating the tourism characteristic index XD is:

[0118] Social behavior data also includes the number of times travel updates are posted.

[0119] It should be noted that the number of travel updates ED refers to the number of times users post travel-related content on various social platforms or travel-related platforms. These contents may include scenic photos during travel, food experience sharing, itinerary records, local cultural introductions, travel insights, etc. The acquisition method is: Many social media platforms provide APIs, allowing developers to obtain their specific data with user authorization, and call these APIs through the AI ​​system to obtain the user's posting records in accordance with the request method and data format specified by the platform, and then filter out travel updates and count the number of releases and calculate their total as the number of travel updates ED.

[0120] Travel data includes travel expenditure amount FA, total travel duration FB and travel playing time FC.

[0121] It should be noted that the travel expense amount FA refers to all expenses spent during the trip, including transportation, accommodation, food, tickets, shopping, etc., which can be obtained by checking the credit card bills and transaction records of electronic payment platforms such as Alipay and WeChat Pay, screening out the consumption records during the trip and summarizing them as the travel expense amount FA. The total travel time FB refers to the sum of all the time spans experienced by each trip from the departure time of the trip to the time when the trip ends and the trip returns to the departure place. The method of obtaining it is: according to the departure time and arrival time on transportation tickets such as air tickets, train tickets, and long-distance bus tickets, the time of each trip is calculated and summarized to obtain the total travel time FB. The travel play time FC refers to the total time that tourists actually spend playing, visiting attractions, participating in entertainment activities, etc. at the tourist destination during the entire trip. The method of obtaining it is: the positioning function of the smartphone can record the user's location changes, and the user's stay time at each attraction can be checked and summarized through the location recording function of the mobile phone system, which is used as the travel play time FC.

[0122] The tourism characteristic index XD is calculated based on the number of travel dynamics releases ED and travel data, and the formula is as follows:

[0123]

[0124] Among them, α 1 is the weight coefficient of tourism consumption efficiency index, ranging from 0.3 to 0.7, and is determined according to the importance of tourism consumption efficiency index to tourism characteristic index XD. 2 is the weight coefficient of the tourism activity index, ranging from 0.3 to 0.7, which is determined according to the importance of the tourism activity index to the tourism characteristic index XD, and α 1 +α 2 =1.

[0125] It should be noted that the operating principle of this formula is: Used to measure the user's travel consumption efficiency. It is used to measure the user's travel activity. By comprehensively considering the influence of tourism consumption efficiency and tourism activity, the tourism characteristic index XD is obtained. The tourism characteristic index XD reflects the user's travel preferences. When the user can have a longer travel time under reasonable consumption and actively publish travel dynamics, the tourism characteristic index XD will be higher.

[0126] Step 206: The method for calculating the user's friendship preference index X is:

[0127] The user friendship preference index X is calculated based on the personality behavior consistency index XA, the life financial deviation index XB, the user friend closeness XC and the travel characteristic index XD. The formula is:

[0128]

[0129] Among them, β 1 is the weight coefficient of the personality behavior consistency index XA, which ranges from 0.2 to 0.4 and is determined according to the influence of the personality behavior consistency index XA on the user's friendship preference index X. 2 is the weight coefficient of the life and finance deviation index XB, which ranges from 0.2 to 0.3 and is determined according to the influence of the life and finance deviation index XB on the user's friendship preference index X. 3 is the weight coefficient of the user's friend closeness XC, ranging from 0.2 to 0.3, and is determined according to the influence of the user's friend closeness XC on the user's friendship preference index X. 4 is the weight coefficient of the tourism characteristic index XD, which ranges from 0.3 to 0.4 and is determined according to the influence of the tourism characteristic index XD on the user's friendship preference index X. 1 +β 2 +β 3 +β 4 =1.

[0130] It should be noted that the operating principle of this formula is: first, the relationship between the four indicators is considered through the geometric mean, then the importance of each indicator is considered through the weighted average, and finally the denominator is scaled to obtain the user's friendship preference index X. The user's friendship preference index X reflects the user's preferences and characteristics in the process of making friends.

[0131] Combined with the contents of step 201 to step 206:

[0132] Calculating the personality behavior consistency index XA based on personality type data and social behavior data can accurately grasp the behavioral pattern tendencies of users based on their personality traits in interpersonal communication. Calculating the life and financial deviation index XB based on life data and financial data can present the coordination status of the user's financial health in daily life and consumption and life rhythm. Calculating the user's friend closeness XC based on social behavior data can clearly understand the depth and closeness of the user's relationship in the existing social network. Calculating the travel characteristic index XD based on social behavior data and travel data is helpful to discover the user's social needs in travel scenarios and enhance the richness and fun of making friends. Calculating the user's friendship preference index X based on the personality behavior consistency index XA, the life and financial deviation index XB, the user's friend closeness XC and the travel characteristic index XD can provide a more comprehensive basis for the friendship recommendation system, reduce invalid social interactions caused by blind recommendations, and improve the success rate of friendship matching.

[0133] Step 3: Calculate the friendship preference index X of other users according to the method in steps 1-2h , based on user friendship preference index X and other user friendship preference index X h Calculate the similarity Y of friendship preferences h .

[0134] Step 3 includes the following:

[0135] Step 301: Calculate the friendship preference similarity Y h The method is:

[0136] Based on user friendship preference index X and other user friendship preference index X h Calculate the similarity Y of friendship preferences h , the formula based on is:

[0137]

[0138] Among them, X h is the user friendship preference index of the hth user on the platform, h is the serial number corresponding to different users, and its value is a positive integer; Y h is the similarity of the user's friendship preference with the hth user on the platform.

[0139] It should be noted that the operating principle of this formula is: It is used to measure the difference between the friendship preference index of a user and other users. The larger the value, the higher the difference. h The lower it is, the smaller the probability of successful dating matching.

[0140] Combined with the content of step 301:

[0141] Based on user friendship preference index X and other user friendship preference index X h Calculate the similarity Y of friendship preferences h , which can accurately match users with similar interests and values, reducing the number of friendship failures caused by lack of common topics or differences in ideas.

[0142] Step 4: Preset a threshold set of similarity levels of friendship preferences; h The friendship preference similarity level is determined by the friendship preference similarity threshold set, and the friend recommendation priority is determined according to the friendship preference similarity level.

[0143] Step 4 includes the following:

[0144] Step 401: The method for determining the similarity level of friendship preferences is:

[0145] The friendship preference similarity level threshold set includes a high similarity threshold YG and a low similarity threshold YD.

[0146] It should be noted that the method for determining the high similarity threshold YG and the low similarity threshold YD is: statistics on user data of successful historical dating matches, and calculate the similarity of dating preferences between successfully matched users according to the above method, screen out the higher 20% of the data and calculate its average value as the reference value of the high similarity threshold YG, and screen out the lower 20% of the data and calculate its average value as the reference value of the low similarity threshold YD.

[0147] According to the similarity of friendship preference Y h The formula for judging the similarity level of friendship preference based on the threshold set of friendship preference similarity level is as follows:

[0148]

[0149] The priority of friend recommendation is determined based on the similarity level of friendship preferences, specifically: high similarity level > medium similarity level > low similarity level.

[0150] Combined with the content of step 401:

[0151] According to the similarity of friendship preference Y h Judging the similarity level of dating preferences by the dating preference similarity threshold set and determining the friend recommendation priority according to the dating preference similarity level can more reasonably allocate platform resources and help the platform better understand user needs and behavior patterns, thereby optimizing the platform's operating strategy.

[0152] See also Figure 3 On the other hand, the present invention also discloses an AI-based dating data analysis and recommendation system, which is used to implement the above-mentioned AI-based dating data analysis and recommendation method, including:

[0153] Data collection module, which can collect user's personality type data, social behavior data, life data, financial data and travel data;

[0154] The index calculation module can calculate the personality behavior consistency index XA based on personality type data and social behavior data; calculate the life balance index BA based on life data, calculate the consumption adaptation life rhythm index BB based on life data and financial data, and calculate the life finance deviation index XB based on the life balance index BA and the consumption adaptation life rhythm index BB; calculate the user's friend closeness XC based on social behavior data; calculate the travel characteristic index XD based on social behavior data and travel data; calculate the user's friendship preference index X based on the personality behavior consistency index XA, the life finance deviation index XB, the user's friend closeness XC and the travel characteristic index XD;

[0155] The friendship preference similarity calculation module can calculate other users' friendship preference index X according to the above method. h, based on user friendship preference index X and other user friendship preference index X h Calculate the similarity Y of friendship preferences h ;

[0156] The recommendation module can preset a threshold set of similarity levels of friendship preferences; according to the friendship preference similarity Y h The friendship preference similarity level is determined by the friendship preference similarity threshold set, and the friend recommendation priority is determined according to the friendship preference similarity level.

[0157] The above embodiments may be implemented in whole or in part by software, hardware, firmware or any other combination thereof. When implemented using software, the above embodiments may be implemented in whole or in part in the form of a computer program product. A person of ordinary skill in the art may appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein may be implemented in electronic hardware or in combination with 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.

[0158] 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.

[0159] The above description is only a specific implementation manner of the present application, but the protection scope of the present application is not limited thereto. Any technician familiar with the 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.

Claims

1. A dating data analysis and recommendation method based on AI, characterized by: include: Step 1: Collect the user's personality type data, social behavior data, life data, financial data and travel data; Step 2: Calculate the personality behavior consistency index XA based on the personality type data and the social behavior data; calculate the life balance index BA based on the life data, calculate the consumption adaptation life rhythm index BB based on the life data and the financial data, and calculate the life finance deviation index XB based on the life balance index BA and the consumption adaptation life rhythm index BB; calculate the user's friend closeness XC based on the social behavior data; calculate the travel characteristic index XD based on the social behavior data and the travel data; calculate the user's friendship preference index X based on the personality behavior consistency index XA, the life finance deviation index XB, the user's friend closeness XC and the travel characteristic index XD; Step 3: Calculate the friendship preference index X of other users according to the method in steps 1-2 h , based on user friendship preference index X and other user friendship preference index X h Calculate the similarity Y of friendship preferences h ; Step 4: Preset a threshold set of similarity levels of friendship preferences; h The friendship preference similarity level is determined by the friendship preference similarity threshold set, and the friend recommendation priority is determined according to the friendship preference similarity level.

2. The AI-based dating data analysis and recommendation method according to claim 1, characterized in that: The method for calculating the personality behavior consistency index XA is: The personality type data include personality expected behavior value A1; Social behavior data includes the number of chats initiated by users within a unit of time A2, the total number of chats A3, the number of offline activities participated in A4, and the total number of activities pushed to users by the platform A5; The personality behavior consistency index XA is calculated based on personality type data and social behavior data, and the formula is:

3. The AI-based dating data analysis and recommendation method according to claim 2, characterized in that: The method for calculating the life balance index BA and the consumption adaptation life rhythm index BB is as follows: Life data including standard deviation of wake-up time Standard deviation of sleep time Standard deviation of work and study time ratio Standard deviation of leisure and entertainment time Standard deviation of time spent on family affairs Standard deviation of social time The frequency score of work and study activities is B7, the frequency score of leisure and entertainment activities is B8, the frequency score of family activities is B9, and the frequency score of social activities is B 10 ; Financial data includes consumption amount D a ; Based on the standard deviation of wake-up time Standard deviation of sleep time Standard deviation of work and study time ratio Standard deviation of leisure and entertainment time Standard deviation of time spent on family affairs and social time percentage standard deviation The calculation formula for life balance index BA is: According to the frequency score of work and study activities B7, the frequency score of leisure and entertainment activities B8, the frequency score of family activities B9, and the frequency score of social activities B 10 and consumption amount D a The formula for calculating the consumption adaptation life rhythm index BB is as follows: Among them, D a is the amount of money spent by the user in the ath consumption category, a is the serial number corresponding to different consumption categories, and its value is [1, b]; b is the total number of consumption categories, and its value is a positive integer.

4. The AI-based friendship data analysis and recommendation method according to claim 3, characterized in that: The method for calculating the life financial deviation index XB is: The life financial deviation index XB is calculated based on the life balance index BA and the consumption adaptation life rhythm index BB, and the formula is as follows: Among them, BA max The maximum value of all user life balance indicators; BA min The minimum value of all user life balance indicators; BB max The maximum value of the life rhythm index for all users; BB min The minimum value of the life rhythm index for all users' consumption adaptation.

5. The AI-based dating data analysis and recommendation method according to claim 4, characterized in that: The method for calculating the user's friend closeness XC is: Social behavior data also includes the number of friends EA d , Number of users EB d And the number of user friend nodes EC d ; EA based on the number of friends d , Number of users EB d And the number of user friend nodes EC d The formula for calculating the user's friend closeness XC is: Among them, EA d is the number of friends of the user on the dth social software, d is the number of different social software, and its value is [1, i]; i is the total number of social software used by the user, and its value is a positive integer; EB d is the number of users using the dth social software, which is a positive integer; EC d is the number of user friend nodes in the dth social software.

6. The AI-based dating data analysis and recommendation method according to claim 5, characterized in that: The method for calculating the tourism characteristic index XD is: Social behavior data also includes the number of travel updates published ED; Travel data includes travel expenditure amount FA, total travel time FB and travel play time FC; The tourism characteristic index XD is calculated based on the number of travel dynamics releases ED and travel data, and the formula is as follows: Among them, α1 is the weight coefficient of the tourism consumption efficiency index, and its value is 0.3~0.7; α2 is the weight coefficient of the tourism activity index, and its value is 0.3~0.7; and α1+α2=1.

7. The AI-based dating data analysis and recommendation method according to claim 6, characterized in that: The method for calculating the user's friendship preference index X is: The user friendship preference index X is calculated based on the personality behavior consistency index XA, the life financial deviation index XB, the user friend closeness XC and the travel characteristic index XD. The formula is: Among them, β1 is the weight coefficient of the personality behavior consistency index XA, and its value is 0.2~0.4; β2 is the weight coefficient of the life financial deviation index XB, and its value is 0.2~0.3; β3 is the weight coefficient of the user friend closeness XC, and its value is 0.2~0.3; β4 is the weight coefficient of the travel characteristics index XD, and its value is 0.3~0.4; and β1+β2+β3+β4=1.

8. The AI-based dating data analysis and recommendation method according to claim 7, characterized in that: Calculate the similarity Y of friendship preferences h The method is: Based on user friendship preference index X and other user friendship preference index X h Calculate the similarity Y of friendship preferences h , the formula based on is: Among them, X h is the user friendship preference index of the hth user on the platform, h is the serial number corresponding to different users, and its value is a positive integer; Y h is the similarity of the user's friendship preference with the hth user on the platform.

9. The AI-based dating data analysis and recommendation method according to claim 8, characterized in that: The method for determining the similarity level of friendship preferences is: The friendship preference similarity level threshold set includes a high similarity threshold YG and a low similarity threshold YD; According to the similarity of friendship preference Y h The formula for judging the similarity level of friendship preference based on the threshold set of friendship preference similarity level is as follows: The priority of friend recommendation is determined according to the similarity level of friendship preferences, specifically: high similarity level > medium similarity level > low similarity level.

10. An AI-based dating data analysis and recommendation system, characterized by: include: Data collection module, which can collect user's personality type data, social behavior data, life data, financial data and travel data; The index calculation module can calculate the personality behavior consistency index XA based on personality type data and social behavior data; calculate the life balance index BA based on life data, calculate the consumption adaptation life rhythm index BB based on life data and financial data, and calculate the life finance deviation index XB based on the life balance index BA and the consumption adaptation life rhythm index BB; calculate the user's friend closeness XC based on social behavior data; calculate the travel characteristic index XD based on social behavior data and travel data; calculate the user's friendship preference index X based on the personality behavior consistency index XA, the life finance deviation index XB, the user's friend closeness XC and the travel characteristic index XD; The friendship preference similarity calculation module can calculate other users' friendship preference index X according to the above method. h , based on user friendship preference index X and other user friendship preference index X h Calculate the similarity Y of friendship preferences h ; The recommendation module can preset a threshold set of similarity levels of friendship preferences; according to the friendship preference similarity Y h The friendship preference similarity level is determined by the friendship preference similarity threshold set, and the friend recommendation priority is determined according to the friendship preference similarity level.

Images