Relationship affinity analysis method and device, computer equipment and storage medium

Abstract

The invention relates to a relationship closeness analysis method and device, computer equipment and a storage medium. The method comprises the steps of obtaining multiple pieces of relationship data of a first involved person and a second involved person; based on multiple set dimensions, relation characteristic indexes of the dimensions are extracted from the multiple pieces of relation data, and the multiple dimensions comprise the communication behavior dimension, the involved person friend dimension, the call behavior dimension, the fund exchange behavior dimension and the position sending behavior dimension; the relation characteristic indexes of each dimension comprise a communication behavior index, a friend characteristic index, a voice call behavior index, a fund exchange behavior index and a position sending behavior index; and inputting the relation characteristic indexes of each dimension into the trained classifier model to obtain the relation closeness output by the trained classifier model. According to the method, the real relation network condition between the involved persons can be reflected, and the efficiency and accuracy of case research and judgment are improved.

Description

Technical Field

[0001] The present application relates to the field of data processing technology, and in particular to a relationship closeness analysis method, apparatus, computer equipment, and storage medium. Background Art

[0002] During the case analysis process, by analyzing the relationship data generated by the communication records, financial transactions, social interactions, etc. between the people involved, the criminal chain between the people involved can be identified, providing strong clues for solving the case. However, the closeness of these relationships generated from communication records, financial transactions, social interactions, etc. cannot be known.

[0003] The current common technical solution to this technical problem is to simply determine the closeness of a relationship by dividing it into relationship types. In the absence of an accurate model to calculate the closeness of the relationship, some non-close relationship data can easily lead to misleading and misjudgment in case investigations. Summary of the Invention

[0004] Based on this, it is necessary to provide a relationship closeness analysis method, device, computer equipment and storage medium to address the above technical problems, which can automatically analyze the closeness of the relationship between the persons involved in the case, provide a basis for judging the credibility of the relationship between the persons involved in the case and whether it is applicable in the case scenario during the case investigation process, and reflect the real relationship network of the persons involved in the case, provide favorable clues for case analysis and judgment, and improve the efficiency and accuracy of case analysis and judgment.

[0005] A relationship closeness analysis method includes: obtaining multiple relationship data between a first person involved in a case and a second person involved in the case; extracting relationship feature indicators of each dimension from the multiple relationship data based on multiple set dimensions, the multiple dimensions including a communication behavior dimension, a friend dimension of the person involved in the case, a call behavior dimension, a fund transaction behavior dimension, and a location sending behavior dimension, and the relationship feature indicators of each dimension including a communication behavior indicator, a friend feature indicator, a voice call behavior indicator, a fund transaction behavior indicator, and a location sending behavior indicator; inputting the relationship feature indicators of each dimension into a trained classifier model to obtain the relationship closeness output by the trained classifier model, wherein the trained classifier model is obtained by model training through sample relationship feature indicators and the relationship closeness between the persons involved in the case corresponding to the sample relationship feature indicators.

[0006] In one embodiment, the communication behavior dimension includes a communication activity dimension, a contact density dimension, and a contact frequency dimension; based on the set multiple dimensions, relationship characteristic indicators of each dimension are extracted from multiple relationship data, including: obtaining the communication time of the first person involved in the case and the second person involved in the case from multiple relationship data, and determining the behavior indicator of the communication activity dimension based on the collection time and communication time of the multiple relationship data; obtaining the number of days between the first and last communication between the first person involved in the case and the second person involved in the case from multiple relationship data, and determining the behavior indicator of the contact density dimension based on the number of days; obtaining the average communication frequency of the first person involved in the case and the second person involved in the case from multiple relationship data, and determining the behavior indicator of the contact frequency dimension based on the average communication frequency.

[0007] In one embodiment, the communication activity dimension includes a recent communication activity dimension, a mid-term communication activity dimension, and a long-term communication activity dimension, and the behavior index of the communication activity dimension is determined based on the collection time and the communication time of multiple relationship data, including: when the difference between the collection time and the communication time is less than or equal to the first time threshold, the behavior index of the recent communication activity dimension is determined; when the difference between the collection time and the communication time is greater than the first time threshold and less than or equal to the second time threshold, the behavior index of the mid-term communication activity dimension is determined; when the difference between the collection time and the communication time is greater than the second time threshold, the behavior index of the long-term communication activity dimension is determined; and / or, the contact density dimension includes an occasional communication dimension, a regular contact dimension, and a close contact dimension, and the behavior index of the contact density dimension is determined based on the interval days, including: when the number of contact days in the interval days is less than the first set number of days, or the number of contact days in the interval days is less than the product of the time interval days and the first set value, the behavior index of the occasional communication dimension is determined; when the number of contact days in the interval days is less than the second set number of days, and the interval When the number of contact days in the number of days is greater than the product of the number of time interval days and a first set value and less than the product of the number of time interval days and a second set value, a behavior indicator of a regular contact dimension is determined; when the number of contact days in the number of days is less than the second set number of days and the number of contact days in the number of days is greater than the product of the number of time interval days and the second set value, a behavior indicator of a close contact dimension is determined; wherein the first set value is greater than 0 and less than 1, the second set value is greater than the first set value and less than 1, and the second set number of days is greater than the first set number of days; and / or, the contact frequency dimension includes a low-frequency contact dimension, a medium-frequency contact dimension, and a high-frequency contact dimension, and determining the behavior indicator of the contact frequency dimension based on the average contact frequency includes: determining the behavior indicator of the low-frequency contact dimension when the average contact frequency is less than the first set frequency, determining the behavior indicator of the medium-frequency contact dimension when the average contact frequency is greater than or equal to the first set frequency and less than the second set frequency, and determining the behavior indicator of the high-frequency contact dimension when the average contact frequency is greater than or equal to the second set frequency; wherein the first set frequency is less than the second set frequency.

[0008] In one embodiment, the communication behavior dimension also includes a communication dimension in a specific time period and a social circle interaction behavior dimension; based on the set multiple dimensions, relationship feature indicators of each dimension are extracted from multiple relationship data, including: obtaining the communication data of the first person involved in the case and the second person involved in the case in a specific time period from multiple relationship data, and determining the behavior indicators of the communication dimension in the specific time period based on the communication data in the specific time period; obtaining the communication data of the social circle interaction behavior of the first person involved in the case and the second person involved in the case from multiple relationship data, and determining the behavior indicators of the social circle interaction behavior dimension based on the communication data of the social circle interaction behavior.

[0009] In one embodiment, relationship characteristic indicators of each dimension are extracted from multiple relationship data based on multiple set dimensions, including: extracting the number of common friends and the number of people belonging to the same community of the first person involved in the case and the second person involved in the case from multiple relationship data; determining the relationship characteristic indicators of the friend dimension of the persons involved in the case based on the number of common friends and the number of people belonging to the same community.

[0010] In one embodiment, relational characteristic indicators of each dimension are extracted from multiple relational data based on multiple set dimensions, including: extracting one or more of the voice call data, video call data, mobile phone call data, call frequency data, and call duration data of the first person involved in the case and the second person involved in the case from multiple relational data; determining the relational characteristic indicators of the call behavior dimension based on one or more of the voice call data, video call data, mobile phone call data, call frequency data, and call duration data.

[0011] In one embodiment, relationship characteristic indicators of each dimension are extracted from multiple relationship data based on set multiple dimensions, including: extracting one or more of the frequency of fund transactions, number of days for fund transactions, fund transaction activity, fixed fund transaction data, fund transaction pattern data, and large-value resource transaction data of the first person involved in the case and the second person involved in the case from multiple relationship data; determining the relationship characteristic indicators of the fund transaction behavior dimension based on one or more of the frequency of fund transactions, number of days for fund transactions, fund transaction activity, fixed fund transaction data, fund transaction pattern data, and large-value resource transaction data.

[0012] A relationship closeness analysis device includes: an acquisition module for acquiring multiple relationship data between a first person involved in a case and a second person involved in the case; an extraction module for extracting relationship feature indicators of each dimension from the multiple relationship data based on multiple set dimensions, the multiple dimensions including a communication behavior dimension, a friend dimension of the person involved in the case, a call behavior dimension, a fund transaction behavior dimension, and a location sending behavior dimension, and the relationship feature indicators of each dimension include a communication behavior indicator, a friend feature indicator, a voice call behavior indicator, a fund transaction behavior indicator, and a location sending behavior indicator; an analysis module for inputting the relationship feature indicators of each dimension into a trained classifier model to obtain the relationship closeness output by the trained classifier model, wherein the trained classifier model is obtained by model training through sample relationship feature indicators and the relationship closeness between the persons involved in the case corresponding to the sample relationship feature indicators.

[0013] A computer device includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the steps of any of the above-mentioned embodiments are implemented.

[0014] A computer-readable storage medium stores a computer program, which implements the steps of any of the above-mentioned method embodiments when executed by a processor.

[0015] The above-mentioned relationship closeness analysis method, device, computer equipment and storage medium obtain multiple relationship data between the first person involved in the case and the second person involved in the case; based on the set multiple dimensions, the relationship feature indicators of each dimension are extracted from the multiple relationship data, the multiple dimensions include the communication behavior dimension, the friend dimension of the person involved in the case, the call behavior dimension, the fund transaction behavior dimension and the location sending behavior dimension, and the relationship feature indicators of each dimension include the communication behavior indicator, the friend feature indicator, the voice call behavior indicator, the fund transaction behavior indicator and the location sending behavior indicator; the relationship feature indicators of each dimension are input into the trained classifier model to obtain the relationship closeness output by the trained classifier model, wherein the trained classifier model is obtained by model training through the sample relationship feature indicators and the relationship closeness between the persons involved in the case corresponding to the sample relationship feature indicators.

[0016] Therefore, it is possible to automatically analyze data such as communication records, financial transactions, social interactions, etc. between people involved in the case, identify key members and potential criminal chains, quickly lock the connection between people involved, victims and witnesses during case investigations, discover implicit patterns and clues in the case, and reflect the real relationship network of people involved, providing favorable clues for case analysis and improving the efficiency and accuracy of case analysis. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1A diagram illustrating an application environment of a relationship closeness analysis method in one embodiment;

[0018] Figure 2 1 is a flow chart of a relationship closeness analysis method in one embodiment;

[0019] Figure 3 is a flowchart of a relationship closeness analysis method in a specific example;

[0020] Figure 4 is a structural block diagram of a relationship closeness analysis device in one embodiment;

[0021] Figure 5 FIG. 1 is a diagram showing the internal structure of a computer device in one embodiment. DETAILED DESCRIPTION

[0022] In order to make the purpose, technical solutions and advantages of this application more clear, the following further describes this application in detail with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain this application and are not intended to limit this application.

[0023] This application provides a relationship closeness analysis method that is applied to Figure 1 In the application environment shown. Among them, the relationship closeness analysis platform 102 is used to execute a relationship closeness analysis method of the present application. Specifically, the relationship closeness analysis platform 102 communicates with the public security system 104 through the network, and obtains multiple relationship data of the first person involved in the case and the second person involved in the case from the public security system 104; based on the set multiple dimensions, the relationship feature indicators of each dimension are extracted from the multiple relationship data, and the multiple dimensions include the communication behavior dimension, the friend dimension of the person involved in the case, the call behavior dimension, the fund transaction behavior dimension and the location sending behavior dimension. The relationship feature indicators of each dimension include the communication behavior indicator, the friend feature indicator, the voice call behavior indicator, the fund transaction behavior indicator and the location sending behavior indicator; the relationship feature indicators of each dimension are input into the trained classifier model to obtain the relationship closeness output by the trained classifier model, wherein the trained classifier model is obtained by model training through the sample relationship feature indicator and the relationship closeness between the persons involved in the case corresponding to the sample relationship feature indicator. Among them, the relationship closeness analysis platform 102 can be implemented by an independent server or a server cluster composed of multiple servers.

[0024] In one embodiment, Figure 2 As shown in the figure, a relationship closeness analysis method is provided, which is applied to Figure 1 Taking the relationship closeness analysis platform 102 in FIG. 1 as an example, the following steps are included:

[0025] S202: Acquire multiple relationship data between the first person involved in the case and the second person involved in the case.

[0026] In this embodiment, the first person involved in the case and the second person involved in the case may be persons involved in the same suspected criminal case. In this application, multiple relationship data of the first person involved in the case and the second person involved in the case are obtained for correlation analysis, which is mainly used by relevant departments to analyze the data of the persons involved in the relevant case, and provide a reliable basis for judging the closeness of the relationship between the persons involved in the case during the case investigation process.

[0027] For example, multiple relational data include relational data generated by communication records, financial transactions, and social interactions between people involved in the case. Through multiple relational data, the criminal chain between people involved in the case can be identified, providing strong clues for solving the case.

[0028] S204, extracting relationship feature indicators of each dimension from multiple relationship data based on the set multiple dimensions, the multiple dimensions including the communication behavior dimension, the friend dimension of the persons involved in the case, the call behavior dimension, the fund transaction behavior dimension, and the location sending behavior dimension, and the relationship feature indicators of each dimension including the communication behavior indicator, the friend feature indicator, the voice call behavior indicator, the fund transaction behavior indicator, and the location sending behavior indicator.

[0029] In this embodiment, dimensions of communication behavior, friends of persons involved in the case, call behavior, funds transaction behavior, and location sending behavior are set, and multiple relationship data are extracted and analyzed through each dimension to obtain relationship feature indicators of each dimension.

[0030] In one embodiment, the communication behavior dimension includes a communication activity dimension, a contact density dimension, and a contact frequency dimension; the above-mentioned extraction of relationship characteristic indicators of each dimension from multiple relationship data based on the set multiple dimensions includes: obtaining the communication time of the first person involved in the case and the second person involved in the case from multiple relationship data, and determining the behavior indicator of the communication activity dimension based on the collection time and communication time of the multiple relationship data; obtaining the number of days between the first and last communication between the first person involved in the case and the second person involved in the case from multiple relationship data, and determining the behavior indicator of the contact density dimension based on the number of days; obtaining the average communication frequency of the first person involved in the case and the second person involved in the case from multiple relationship data, and determining the behavior indicator of the contact frequency dimension based on the average communication frequency.

[0031] In one example, the communication activity dimension includes a recent communication activity dimension, a mid-term communication activity dimension, and a long-term communication activity dimension. The above-mentioned determination of the behavior indicator of the communication activity dimension based on the collection time and the communication time of the plurality of relationship data includes: when the difference between the collection time and the communication time is less than or equal to the first time threshold, determining the behavior indicator of the recent communication activity dimension; when the difference between the collection time and the communication time is greater than the first time threshold and less than or equal to the second time threshold, determining the behavior indicator of the mid-term communication activity dimension; when the difference between the collection time and the communication time is greater than the second time threshold, determining the behavior indicator of the long-term communication activity dimension; and / or, the contact density dimension includes an occasional communication dimension, a regular contact dimension, and a close contact dimension. The above-mentioned determination of the behavior indicator of the contact density dimension based on the interval days includes: when the number of contact days in the interval days is less than the first set number of days, or the number of contact days in the interval days is less than the product of the time interval days and the first set value, then determining the behavior indicator of the occasional communication dimension; when the number of contact days in the interval days is less than the second set number of days, and the interval days When the number of contact days in the number is greater than the product of the time interval days and the first set value and less than the product of the time interval days and the second set value, a behavior indicator of the regular contact dimension is determined; when the number of contact days in the interval days is less than the second set number of days and the number of contact days in the interval days is greater than the product of the time interval days and the second set value, a behavior indicator of the close contact dimension is determined; wherein the first set value is greater than 0 and less than 1, the second set value is greater than the first set value and less than 1, and the second set number of days is greater than the first set number of days; and / or, the contact frequency dimension includes a low-frequency contact dimension, a medium-frequency contact dimension, and a high-frequency contact dimension, and the above-mentioned determination of the behavior indicator of the contact frequency dimension based on the average contact frequency includes: determining the behavior indicator of the low-frequency contact dimension when the average contact frequency is less than the first set frequency, determining the behavior indicator of the medium-frequency contact dimension when the average contact frequency is greater than or equal to the first set frequency and less than the second set frequency, and determining the behavior indicator of the high-frequency contact dimension when the average contact frequency is greater than or equal to the second set frequency; wherein the first set frequency is less than the second set frequency.

[0032] In this embodiment, the extraction of relationship feature indicators of the communication behavior dimension includes:

[0033] 1) Communication is active recently

[0034] This period refers to the contact period that is relatively short from the time of collection. Within three months, it is determined whether the following conditions are met:

[0035] Collection time - communication time <= 90 days.

[0036] 2) Active communication in the medium term

[0037] This period covers the period from the end of the recent contact to one year. The mid-term contact period is from three months ago to one year, and is determined by whether the following conditions are met:

[0038] 90 days < collection time - communication time <= 365 days.

[0039] 3) Long-term communication activity

[0040] This period refers to contact records from more than one year ago, and determines whether the following conditions are met:

[0041] Collection time - communication time > 365 days.

[0042] 4) Occasional communication

[0043] The interval is the number of days between the first and last contact. Let y be the number of contact days in the interval, and y1 be the time difference between the last contact time in the interval and the start time of the interval. If the following conditions are met, it is considered occasional contact:

[0044] y<1 / 4(y1) or y<10 days.

[0045] 5) Regular Contact

[0046] The interval is the number of days between the first and last contact. Let y be the number of contact days in the interval, and y1 be the time difference between the last contact time in the interval and the start time of the interval. If the following conditions are met, it is considered regular contact:

[0047] 1 / 2(y1)>y>1 / 4(y1) and y>22 days.

[0048] 6) Close contact

[0049] The interval is the number of days between the first and last contact. Let y be the number of contact days in the interval, and y1 be the time difference between the last contact time in the interval and the start time of the interval. If the following conditions are met, it is considered a close connection:

[0050] y>1 / 2(y1) and y>22 days.

[0051] 7) Low-frequency contact

[0052] Average contact frequency = total number of contacts / number of days with contact. Low-frequency contact is defined as meeting the following conditions:

[0053] The average communication frequency is <3 times / day.

[0054] 8) IF contact

[0055] Average contact frequency = total number of contacts / number of days with contact. Medium-frequency contact is defined as meeting the following conditions:

[0056] 3 times / day <= average communication frequency <= 10 times / day.

[0057] 9) High-frequency contact

[0058] Average contact frequency = total number of contacts / number of days with contact. High-frequency contact is defined as meeting the following conditions:

[0059] The average communication frequency is >10 times / day.

[0060] In one embodiment, the communication behavior dimension also includes a communication dimension in a specific time period and a social circle interaction behavior dimension; the above-mentioned extraction of relationship characteristic indicators of each dimension from multiple relationship data based on the set multiple dimensions includes: obtaining the communication data of the first person involved in the case and the second person involved in the case in a specific time period from multiple relationship data, and determining the behavior indicators of the communication dimension in the specific time period based on the communication data in the specific time period; obtaining the communication data of the social circle interaction behavior of the first person involved in the case and the second person involved in the case from multiple relationship data, and determining the behavior indicators of the social circle interaction behavior dimension based on the communication data of the social circle interaction behavior.

[0061] Specifically, specific time periods include time periods in days and time periods in hours. For example, if the persons involved in the case have communication characteristics on holidays, mornings, evenings, and weekends, the specific time periods are shown in the following table:

[0062]

[0063] For example, divide 0-24 o'clock into 24 time periods of 1 hour each, and analyze the communication frequency of the persons involved in the time periods of 0-6, 7-12, 13-18, and 19-23. If the proportion on the same day exceeds 50% and the number of days is >30, it is judged to meet the conditions.

[0064] Specifically, the social circle interaction behavior can be a friend circle interaction behavior, specifically: an interaction behavior with the friend circle of a corresponding friend.

[0065] Therefore, the relationship characteristic indicators of the communication behavior dimension of the first person involved in the case and the second person involved in the case can be extracted through multiple dimensions, providing a reliable basis for judging the credibility of the relationship between the persons involved in the case and whether it is applicable in the case scenario during the case investigation process.

[0066] In one embodiment, relationship characteristic indicators of each dimension are extracted from multiple relationship data based on the set multiple dimensions, including: extracting the number of common friends of the first person involved in the case and the number of people belonging to the same community from the multiple relationship data; determining the relationship characteristic indicators of the friend dimension of the people involved in the case based on the number of common friends and the number of people belonging to the same community.

[0067] Specifically, the extraction of relationship characteristic indicators of the friend dimension of the persons involved in the case includes:

[0068] 1) Mutual Friends I

[0069] 1 <= number of common friends < 5.

[0070] 2) Mutual Friends II

[0071] 5 <= number of common friends < 10.

[0072] 3) Mutual Friends III

[0073] 10 <= number of common friends.

[0074] 4) Friends in the same group I

[0075] 1<=Number of same groups<5.

[0076] 5) Friends in the same group II

[0077] 5<=Number of same groups<10.

[0078] 6) Friends in the same group III

[0079] 10 <= number of the same groups.

[0080] 7) Note friends

[0081] Note friends for the owner.

[0082] 8) Star friends

[0083] Friends are star friends and special attention friends.

[0084] Therefore, the relationship characteristic indicators of the friend dimension of the first person involved in the case and the second person involved in the case can be extracted through multiple dimensions, providing a reliable basis for judging the credibility of the relationship between the persons involved in the case and whether it is applicable in the case scenario during the case investigation process.

[0085] In one embodiment, the above-mentioned extraction of relationship characteristic indicators of each dimension from multiple relationship data based on the set multiple dimensions includes: extracting one or more of the voice call data, video call data, mobile phone call data, call frequency data and call duration data of the first person involved in the case and the second person involved in the case from multiple relationship data; determining the relationship characteristic indicators of the call behavior dimension based on one or more of the voice call data, video call data, mobile phone call data, call frequency data and call duration data.

[0086] Specifically, the extraction of relationship feature indicators of the call behavior dimension includes:

[0087] 1) Voice call

[0088] There are voice calls using apps such as WeChat.

[0089] 2) Video Call

[0090] There are video calls using apps such as WeChat.

[0091] 3) Mobile phone calls

[0092] There are calls made using a mobile phone number.

[0093] 4) High-frequency calls

[0094] The number of calls in a day is greater than or equal to 10 times, or the number of calls in a week is greater than or equal to 30 times.

[0095] 5) Long calls

[0096] A single call lasts more than 5 minutes.

[0097] Therefore, the relationship characteristic indicators of the call behavior dimension of the first person involved in the case and the second person involved in the case can be extracted through multiple dimensions, providing a reliable basis for judging the credibility of the relationship between the persons involved in the case and whether it is applicable in the case scenario during the case investigation process.

[0098] In one embodiment, the above-mentioned extraction of relationship characteristic indicators of each dimension from multiple relationship data based on the set multiple dimensions includes: extracting one or more of the frequency of fund transactions, number of days for fund transactions, fund transaction activity, fixed fund transaction data, fund transaction regularity data, and large-value resource transaction data of the first person involved in the case and the second person involved in the case from multiple relationship data; determining the relationship characteristic indicators of the fund transaction behavior dimension based on one or more of the frequency of fund transactions, number of days for fund transactions, fund transaction activity, fixed fund transaction data, fund transaction regularity data, and large-value resource transaction data.

[0099] Specifically, the extraction of relationship characteristic indicators in the dimension of fund transaction behavior includes:

[0100] 1) Low-frequency capital transactions

[0101] 1 time <= total number of fund transactions < 5 times.

[0102] 2) Medium frequency fund transactions

[0103] 5 times <= Number of fund transactions < 10 times.

[0104] 3) High-frequency capital transactions

[0105] Number of fund transactions >= 10 times.

[0106] 4) Short-term capital transactions

[0107] 1 day <= number of days for fund transactions < 5 days.

[0108] 5) Medium-term capital transactions

[0109] 5 days <= Funds transaction days < 10 days.

[0110] 6) Long-term capital transactions

[0111] 10 days <= number of days for fund transactions.

[0112] 7) Funding transactions have been active recently

[0113] This period refers to the contact period that is relatively short from the time of collection. Within three months, it is determined whether the following conditions are met:

[0114] Collection time - communication time <= 90 days.

[0115] 8) Capital flows are active in the medium term

[0116] This period covers the period from the end of the recent contact to one year. The mid-term contact period is from three months ago to one year, and is determined by whether the following conditions are met:

[0117] 90 days < collection time - communication time <= 365 days.

[0118] 9) Active forward capital flows

[0119] This period refers to contact records from more than one year ago, and determines whether the following conditions are met:

[0120] Collection time - communication time > 365 days.

[0121] 10) High-frequency fixed capital transactions

[0122] The amount of fund transactions is greater than or equal to RMB 200, and the number of fund transactions is greater than or equal to 50 times.

[0123] 11) Regular Fund Transactions - Periodic Frequency

[0124] The number of fund transactions follows a fixed cycle. 7-day, 15-day, and 30-day sliding windows are used to count whether there is any transfer activity within the window. If there is a transfer for three consecutive cycles, the corresponding label will be added.

[0125] 12) Regular Funds Flows - Periodic Limits

[0126] Fund transfer quotas have periodic characteristics. We use 7-day, 15-day, and 30-day sliding windows to count whether the transfer quota falls within a certain interval (1-8 interval levels). If the corresponding quota is met for more than 3 periods, the corresponding label will be added. For example, as shown in the following table:

[0127] interval Amount range (yuan) 1 0-100 2 100-300 3 300-500 4 500-700 5 700-1000 6 1000-1500 7 1500-3000 8 More than 3000

[0128] 13) High-frequency capital transactions

[0129] The number of fund transactions is counted based on four periodic windows: hourly, daily, weekly, and monthly. High-frequency fund transactions are those that meet the following conditions:

[0130] Number of times per hour >= 2, number of times per day >= 3, number of times per week >= 5, number of times per month >= 7.

[0131] 14) Large-amount capital transactions

[0132] Fund transactions exceeding RMB 3,000 are considered large-value fund transactions.

[0133] Therefore, it is possible to extract the relationship characteristic indicators of the financial transaction behavior dimension of the first person involved in the case and the second person involved in the case through multiple dimensions, providing a reliable basis for judging the credibility of the relationship between the persons involved in the case and whether it is applicable in the case scenario during the case investigation process.

[0134] In one embodiment, the above-mentioned extraction of relationship characteristic indicators of each dimension from multiple relationship data based on the set multiple dimensions includes: extracting the location sending behavior data of the first person involved in the case and the second person involved in the case from multiple relationship data; determining the relationship characteristic indicators of the location sending behavior dimension based on the location sending behavior data; and / or, extracting the friend message collection behavior data of the first person involved in the case and the second person involved in the case from multiple relationship data; determining the relationship characteristic indicators of the friend message collection behavior dimension based on the friend message collection behavior data.

[0135] Specifically, the extraction of relationship feature indicators of the location sending behavior dimension includes:

[0136] 1) Send location behavior

[0137] There is an action to send the location.

[0138] 2) High-frequency position transmission

[0139] High-frequency location sharing behavior in a certain period determines whether it meets the following conditions:

[0140] The number of times the location is sent in a certain day is greater than 3 times or the number of times the location is sent in a certain week is greater than 10 times.

[0141] Specifically, the extraction of relationship feature indicators in the dimension of collecting friend messages includes:

[0142] There is an act of collecting friend information.

[0143] Therefore, the relationship characteristic indicators of the location sending behavior dimension and the friend message collection behavior dimension of the first person involved in the case and the second person involved in the case can be extracted through multiple dimensions, providing a reliable judgment basis for the credibility of the relationship between the persons involved in the case and whether it is applicable in the case scenario during the case investigation process.

[0144] S206, inputting the relationship feature indicators of each dimension into the trained classifier model to obtain the relationship closeness output by the trained classifier model, wherein the trained classifier model is obtained by model training through the sample relationship feature indicators and the relationship closeness between the persons involved in the case corresponding to the sample relationship feature indicators.

[0145] In this embodiment, the process of calculating the closeness of the relationship between the first person involved in the case and the second person involved in the case is as follows:

[0146] Based on the relationship feature indicators of each dimension extracted above, the relationship closeness is calculated using the XGBoost classifier algorithm.

[0147] XGBoost is an optimized gradient boosting decision tree algorithm widely used in machine learning. It gradually optimizes the objective function by iteratively adding trees, with each new tree adding a new prediction based on the residuals of all previous trees. XGBoost's objective function consists of a loss function and a regularization term, which measures the difference between the model's predictions and the true values ​​and controls model complexity. XGBoost can be used to solve a variety of machine learning problems, including classification, regression, and ranking.

[0148] How it works: XGBoost iteratively adds trees (typically decision trees) to gradually optimize an objective function. Each step attempts to add a new tree that predicts the residual (or error) of all the previous trees. By continuously reducing the residual, the model becomes increasingly powerful. Its objective function consists of two parts: a loss function (which measures the difference between the model's predictions and the true values) and a regularization term (which controls model complexity).

[0149] For gradient-based decision tree models like XGBoost, when performing classification tasks, they train multiple trees to optimize a loss function, which typically involves probabilistic calculations (such as using a logarithmic loss function to optimize binary or multi-classification problems). Each tree attempts to correct the errors of the previous tree, and ultimately, the ensemble of these trees is used to calculate the final prediction probability in a specific way (such as by taking the average).

[0150] The training process of the XGBoost classifier model is as follows:

[0151] 1. Use SAPRK-SQL to load different resource data, clean the loaded data, and remove illegal format data.

[0152] 2. Perform necessary preprocessing on the data, such as filling missing values, data standardization or normalization, etc.

[0153] 3. Divide the data set into training set, validation set and test set.

[0154] 4. Define the model parameters as shown in the following table:

[0155]

[0156]

[0157] 5. Build DMatrix. DMatrix is ​​a special data structure used to store data in XGBoost, which can improve training speed and memory utilization.

[0158] 6. Train the model using the defined parameters.

[0159] The model evaluation of the XGBoost classifier model is as follows:

[0160] 1. Model prediction, use the trained model to make predictions.

[0161] 2. Evaluate model performance and use different evaluation indicators to evaluate the performance of the model, such as accuracy, AUC-ROC, etc.

[0162] 3. Parameter tuning, use grid search GridSearchCV() to find the best parameter combination, and retrain the model using the best parameters.

[0163] Therefore, by extracting multi-dimensional relationship feature indicators and calculating the relationship closeness based on the XGBoost classifier algorithm, accurate relationship closeness calculation results are obtained through multiple model training and parameter tuning, which provides a reliable judgment basis for the credibility of the relationship between the persons involved in the case and whether it is applicable in the case scenario during the case investigation process, and reflects the real relationship network of the persons involved, providing favorable clues for case analysis and improving the efficiency and accuracy of case analysis.

[0164] In one embodiment, after the above-mentioned step S206, that is, after obtaining the relationship closeness output by the trained classifier model, it also includes: obtaining the case information involved in the first person involved in the case and the second person involved in the case; and associating the relationship closeness with the case information for displaying for application analysis of the case involved.

[0165] Specifically, after the result data of relationship closeness is calculated through the trained XGBoost classifier model, it is integrated with the calculation and display of related functional applications such as personnel association, key, intimate, and close relationship analysis, common relationship mining, and social relationship network expansion. The operation interface provides functions such as sorting, filtering, and aggregation statistics by relationship closeness.

[0166] A relationship closeness analysis method based on each of the above embodiments, Figure 3 As shown, a specific example is provided.

[0167] Therefore, by extracting multi-dimensional relationship feature indicators and calculating the relationship closeness based on the XGBoost classifier algorithm, accurate relationship closeness score calculation results are obtained through multiple model training and parameter tuning, which provides a reliable judgment basis for the credibility of the relationship between the persons involved in the case and whether it is applicable in the case scenario during the case investigation process, and reflects the real relationship network of the persons involved, providing favorable clues for case analysis and improving the efficiency and accuracy of case analysis.

[0168] It should be understood that although the steps in the flowchart are shown in sequence as indicated by the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless otherwise specified herein, there is no strict order restriction on the execution of these steps, and these steps can be executed in other orders. Moreover, at least some of the steps in the accompanying drawings may include multiple sub-steps or multiple stages, and these sub-steps or stages are not necessarily executed at the same time, but can be executed at different times. The order of execution of these sub-steps or stages is not necessarily sequential, but can be executed in turn or alternately with other steps or at least a portion of the sub-steps or stages of other steps.

[0169] This application also provides a relationship closeness analysis device. Figure 4As shown, a relationship closeness analysis device includes an acquisition module 402, an extraction module 404, and an analysis module 406. The acquisition module 402 is used to obtain multiple relationship data between a first person involved in the case and a second person involved in the case; the extraction module 404 is used to extract relationship feature indicators of each dimension from the multiple relationship data based on multiple set dimensions, the multiple dimensions including the communication behavior dimension, the friend dimension of the person involved in the case, the call behavior dimension, the fund transaction behavior dimension, and the location sending behavior dimension, and the relationship feature indicators of each dimension include the communication behavior indicator, the friend feature indicator, the voice call behavior indicator, the fund transaction behavior indicator, and the location sending behavior indicator; the analysis module 406 is used to input the relationship feature indicators of each dimension into a trained classifier model to obtain the relationship closeness output by the trained classifier model, wherein the trained classifier model is obtained by model training through sample relationship feature indicators and the relationship closeness between the persons involved in the case corresponding to the sample relationship feature indicators.

[0170] In one embodiment, the communication behavior dimension includes a communication activity dimension, a contact density dimension, and a contact frequency dimension; based on the set multiple dimensions, relationship characteristic indicators of each dimension are extracted from multiple relationship data, including: obtaining the communication time of the first person involved in the case and the second person involved in the case from multiple relationship data, and determining the behavior indicator of the communication activity dimension based on the collection time and communication time of the multiple relationship data; obtaining the number of days between the first and last communication between the first person involved in the case and the second person involved in the case from multiple relationship data, and determining the behavior indicator of the contact density dimension based on the number of days; obtaining the average communication frequency of the first person involved in the case and the second person involved in the case from multiple relationship data, and determining the behavior indicator of the contact frequency dimension based on the average communication frequency.

[0171] In one embodiment, the communication activity dimension includes a recent communication activity dimension, a mid-term communication activity dimension, and a long-term communication activity dimension, and the behavior index of the communication activity dimension is determined based on the collection time and the communication time of multiple relationship data, including: when the difference between the collection time and the communication time is less than or equal to the first time threshold, the behavior index of the recent communication activity dimension is determined; when the difference between the collection time and the communication time is greater than the first time threshold and less than or equal to the second time threshold, the behavior index of the mid-term communication activity dimension is determined; when the difference between the collection time and the communication time is greater than the second time threshold, the behavior index of the long-term communication activity dimension is determined; and / or, the contact density dimension includes an occasional communication dimension, a regular contact dimension, and a close contact dimension, and the behavior index of the contact density dimension is determined based on the interval days, including: when the number of contact days in the interval days is less than the first set number of days, or the number of contact days in the interval days is less than the product of the time interval days and the first set value, the behavior index of the occasional communication dimension is determined; when the number of contact days in the interval days is less than the second set number of days, and the interval When the number of contact days in the number of days is greater than the product of the number of time interval days and a first set value and less than the product of the number of time interval days and a second set value, a behavior indicator of a regular contact dimension is determined; when the number of contact days in the number of days is less than the second set number of days and the number of contact days in the number of days is greater than the product of the number of time interval days and the second set value, a behavior indicator of a close contact dimension is determined; wherein the first set value is greater than 0 and less than 1, the second set value is greater than the first set value and less than 1, and the second set number of days is greater than the first set number of days; and / or, the contact frequency dimension includes a low-frequency contact dimension, a medium-frequency contact dimension, and a high-frequency contact dimension, and determining the behavior indicator of the contact frequency dimension based on the average contact frequency includes: determining the behavior indicator of the low-frequency contact dimension when the average contact frequency is less than the first set frequency, determining the behavior indicator of the medium-frequency contact dimension when the average contact frequency is greater than or equal to the first set frequency and less than the second set frequency, and determining the behavior indicator of the high-frequency contact dimension when the average contact frequency is greater than or equal to the second set frequency; wherein the first set frequency is less than the second set frequency.

[0172] In one embodiment, the communication behavior dimension also includes a communication dimension in a specific time period and a social circle interaction behavior dimension; based on the set multiple dimensions, relationship feature indicators of each dimension are extracted from multiple relationship data, including: obtaining the communication data of the first person involved in the case and the second person involved in the case in a specific time period from multiple relationship data, and determining the behavior indicators of the communication dimension in the specific time period based on the communication data in the specific time period; obtaining the communication data of the social circle interaction behavior of the first person involved in the case and the second person involved in the case from multiple relationship data, and determining the behavior indicators of the social circle interaction behavior dimension based on the communication data of the social circle interaction behavior.

[0173] In one embodiment, relationship characteristic indicators of each dimension are extracted from multiple relationship data based on multiple set dimensions, including: extracting the number of common friends and the number of people belonging to the same community of the first person involved in the case and the second person involved in the case from multiple relationship data; determining the relationship characteristic indicators of the friend dimension of the persons involved in the case based on the number of common friends and the number of people belonging to the same community.

[0174] In one embodiment, relational characteristic indicators of each dimension are extracted from multiple relational data based on multiple set dimensions, including: extracting one or more of the voice call data, video call data, mobile phone call data, call frequency data, and call duration data of the first person involved in the case and the second person involved in the case from multiple relational data; determining the relational characteristic indicators of the call behavior dimension based on one or more of the voice call data, video call data, mobile phone call data, call frequency data, and call duration data.

[0175] In one embodiment, relationship characteristic indicators of each dimension are extracted from multiple relationship data based on set multiple dimensions, including: extracting one or more of the frequency of fund transactions, number of days for fund transactions, fund transaction activity, fixed fund transaction data, fund transaction pattern data, and large-value resource transaction data of the first person involved in the case and the second person involved in the case from multiple relationship data; determining the relationship characteristic indicators of the fund transaction behavior dimension based on one or more of the frequency of fund transactions, number of days for fund transactions, fund transaction activity, fixed fund transaction data, fund transaction pattern data, and large-value resource transaction data.

[0176] For the specific definition of a relationship closeness analysis device, please refer to the definition of a relationship closeness analysis method above, and will not be repeated here. Each module in the above-mentioned relationship closeness analysis device can be implemented in whole or in part by software, hardware, or a combination thereof. The above-mentioned modules can be embedded in or independent of the processor in the computer device in the form of hardware, or can be stored in the memory of the computer device in the form of software, so that the processor can call and execute the operations corresponding to the above modules.

[0177] In one embodiment, a computer device is provided. The computer device may be Figure 1 The internal structure of the server in the relationship closeness analysis platform 102 can be as shown in FIG. Figure 5As shown. The computer device includes a processor, a memory, a network interface and a database connected via a system bus. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the computer device is used to store relationship data of multiple persons involved in the case. The network interface of the computer device is used to communicate with the public security system via a network connection. When the computer program is executed by the processor, a relationship closeness analysis method is implemented.

[0178] In one embodiment, a computer device is provided, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the following steps when executing the computer program: obtaining multiple relationship data between a first person involved in the case and a second person involved in the case; extracting relationship feature indicators of each dimension from the multiple relationship data based on multiple set dimensions, wherein the multiple dimensions include a communication behavior dimension, a friend dimension of the person involved in the case, a call behavior dimension, a fund transaction behavior dimension, and a location sending behavior dimension, and the relationship feature indicators of each dimension include a communication behavior indicator, a friend feature indicator, a voice call behavior indicator, a fund transaction behavior indicator, and a location sending behavior indicator; inputting the relationship feature indicators of each dimension into a trained classifier model to obtain the relationship closeness output by the trained classifier model, wherein the trained classifier model is obtained by model training using sample relationship feature indicators and the relationship closeness between the persons involved in the case corresponding to the sample relationship feature indicators.

[0179] In one embodiment, the communication behavior dimension includes a communication activity dimension, a contact density dimension, and a contact frequency dimension; when the processor executes a computer program to implement the above-mentioned step of extracting the relationship characteristic indicators of each dimension from multiple relationship data based on the set multiple dimensions, the following steps are specifically implemented: obtaining the communication time of the first person involved in the case and the second person involved in the case from multiple relationship data, and determining the behavior indicator of the communication activity dimension based on the collection time and communication time of the multiple relationship data; obtaining the number of days between the first and last communication between the first person involved in the case and the second person involved in the case from multiple relationship data, and determining the behavior indicator of the contact density dimension based on the number of days; obtaining the average communication frequency of the first person involved in the case and the second person involved in the case from multiple relationship data, and determining the behavior indicator of the contact frequency dimension based on the average communication frequency.

[0180] In one embodiment, the communication activity dimension includes a recent communication activity dimension, a medium-term communication activity dimension, and a long-term communication activity dimension. When the processor executes the computer program to implement the above-mentioned step of determining the behavior indicator of the communication activity dimension based on the collection time and the communication time of the plurality of relationship data, the following steps are specifically implemented: when the difference between the collection time and the communication time is less than or equal to the first time threshold, the behavior indicator of the recent communication activity dimension is determined; when the difference between the collection time and the communication time is greater than the first time threshold and less than or equal to the second time threshold, the behavior indicator of the medium-term communication activity dimension is determined; when the difference between the collection time and the communication time is greater than the second time threshold, the behavior indicator of the long-term communication activity dimension is determined; and / or, the contact density dimension includes an occasional communication dimension, a regular contact dimension, and a close contact dimension, and the behavior indicator of the contact density dimension is determined based on the interval days, including: when the number of communication days in the interval days is less than the first set number of days, or when the number of communication days in the interval days is less than the product of the time interval days and the first set value, the behavior indicator of the occasional communication dimension is determined; when the number of communication days in the interval days is less than the first set number of days, or when the number of communication days in the interval days is less than the product of the time interval days and the first set value, the behavior indicator of the occasional communication dimension is determined; When the number of communication days in the interval is less than the second set number of days, and the number of communication days in the interval is greater than the product of the time interval days and the first set value and less than the product of the time interval days and the second set value, a behavior indicator of the regular communication dimension is determined; when the number of communication days in the interval is less than the second set number of days, and the number of communication days in the interval is greater than the product of the time interval days and the second set value, a behavior indicator of the close communication dimension is determined; wherein the first set value is greater than 0 and less than 1, the second set value is greater than the first set value and less than 1, and the second set number of days is greater than the first set value; and / or the contact frequency dimension includes a low-frequency contact dimension, a medium-frequency contact dimension, and a high-frequency contact dimension, and determining the behavior indicator of the contact frequency dimension based on the average communication frequency includes: when the average communication frequency is less than the first set frequency, determining the behavior indicator of the low-frequency contact dimension; when the average communication frequency is greater than or equal to the first set frequency and less than the second set frequency, determining the behavior indicator of the medium-frequency contact dimension; and when the average communication frequency is greater than or equal to the second set frequency, determining the behavior indicator of the high-frequency contact dimension; wherein the first set frequency is less than the second set frequency.

[0181] In one embodiment, the communication behavior dimension also includes a communication dimension in a specific time period and a social circle interaction behavior dimension; when the processor executes a computer program to implement the above-mentioned step of extracting relationship characteristic indicators of each dimension from multiple relationship data based on the set multiple dimensions, the following steps are specifically implemented: obtaining the communication data of the first person involved in the case and the second person involved in the case in a specific time period from multiple relationship data, and determining the behavior indicators of the communication dimension in the specific time period based on the communication data in the specific time period; obtaining the communication data of the social circle interaction behavior of the first person involved in the case and the second person involved in the case from multiple relationship data, and determining the behavior indicators of the social circle interaction behavior dimension based on the communication data of the social circle interaction behavior.

[0182] In one embodiment, when the processor executes a computer program to implement the above-mentioned steps of extracting relationship characteristic indicators of each dimension from multiple relationship data based on the set multiple dimensions, the following steps are specifically implemented: extracting the number of common friends and the number of people belonging to the same community of the first person involved in the case and the second person involved in the case from the multiple relationship data; determining the relationship characteristic indicators of the friend dimension of the persons involved in the case based on the number of common friends and the number of people belonging to the same community.

[0183] In one embodiment, when the processor executes a computer program to implement the above-mentioned step of extracting relationship characteristic indicators of each dimension from multiple relationship data based on the set multiple dimensions, the following steps are specifically implemented: extracting one or more of the voice call data, video call data, mobile phone call data, call frequency data and call duration data of the first person involved in the case and the second person involved in the case from the multiple relationship data; determining the relationship characteristic indicators of the call behavior dimension based on one or more of the voice call data, video call data, mobile phone call data, call frequency data and call duration data.

[0184] In one embodiment, when the processor executes a computer program to implement the above-mentioned step of extracting relationship characteristic indicators of each dimension from multiple relationship data based on the set multiple dimensions, the following steps are specifically implemented: extracting one or more of the frequency of fund transactions, number of days for fund transactions, fund transaction activity, fixed fund transaction data, fund transaction regularity data, and large-value resource transaction data of the first person involved in the case and the second person involved in the case from the multiple relationship data; determining the relationship characteristic indicators of the fund transaction behavior dimension based on one or more of the frequency of fund transactions, number of days for fund transactions, fund transaction activity, fixed fund transaction data, fund transaction regularity data, and large-value resource transaction data.

[0185] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored, and when the computer program is executed by a processor, the following steps are implemented: obtaining multiple relationship data between a first person involved in the case and a second person involved in the case; extracting relationship feature indicators of each dimension from the multiple relationship data based on multiple set dimensions, the multiple dimensions including a communication behavior dimension, a friend dimension of the person involved in the case, a call behavior dimension, a fund transaction behavior dimension, and a location sending behavior dimension, and the relationship feature indicators of each dimension including a communication behavior indicator, a friend feature indicator, a voice call behavior indicator, a fund transaction behavior indicator, and a location sending behavior indicator; inputting the relationship feature indicators of each dimension into a trained classifier model to obtain the relationship closeness output by the trained classifier model, wherein the trained classifier model is obtained by model training through sample relationship feature indicators and the relationship closeness between the persons involved in the case corresponding to the sample relationship feature indicators.

[0186] In one embodiment, the communication behavior dimension includes a communication activity dimension, a contact density dimension, and a contact frequency dimension; when the computer program is executed by the processor to implement the above-mentioned step of extracting the relationship characteristic indicators of each dimension from multiple relationship data based on the set multiple dimensions, the following steps are specifically implemented: obtaining the communication time of the first person involved in the case and the second person involved in the case from multiple relationship data, and determining the behavior indicator of the communication activity dimension based on the collection time and communication time of the multiple relationship data; obtaining the number of days between the first and last communication between the first person involved in the case and the second person involved in the case from multiple relationship data, and determining the behavior indicator of the contact density dimension based on the number of days; obtaining the average communication frequency of the first person involved in the case and the second person involved in the case from multiple relationship data, and determining the behavior indicator of the contact frequency dimension based on the average communication frequency.

[0187] In one embodiment, the communication activity dimension includes a recent communication activity dimension, a medium-term communication activity dimension, and a long-term communication activity dimension. When the computer program is executed by the processor to implement the above-mentioned step of determining the behavior indicator of the communication activity dimension based on the collection time and the communication time of multiple relationship data, the following steps are specifically implemented: when the difference between the collection time and the communication time is less than or equal to the first time threshold, the behavior indicator of the recent communication activity dimension is determined; when the difference between the collection time and the communication time is greater than the first time threshold and less than or equal to the second time threshold, the behavior indicator of the medium-term communication activity dimension is determined; when the difference between the collection time and the communication time is greater than the second time threshold, the behavior indicator of the long-term communication activity dimension is determined; and / or, the contact density dimension includes an occasional communication dimension, a regular contact dimension, and a close contact dimension, and the behavior indicator of the contact density dimension is determined according to the interval days, including: when the number of contact days in the interval days is less than the first set number of days, or when the number of contact days in the interval days is less than the product of the time interval days and the first set value, the behavior indicator of the occasional communication dimension is determined; when the number of contact days in the interval days is less than the first set number of days, or when the number of contact days in the interval days is less than the product of the time interval days and the first set value, the behavior indicator of the occasional communication dimension is determined; When the number of communication days in the interval is less than the second set number of days, and the number of communication days in the interval is greater than the product of the time interval days and the first set value and less than the product of the time interval days and the second set value, a behavior indicator of the regular communication dimension is determined; when the number of communication days in the interval is less than the second set number of days, and the number of communication days in the interval is greater than the product of the time interval days and the second set value, a behavior indicator of the close communication dimension is determined; wherein the first set value is greater than 0 and less than 1, the second set value is greater than the first set value and less than 1, and the second set number of days is greater than the first set value; and / or the contact frequency dimension includes a low-frequency contact dimension, a medium-frequency contact dimension, and a high-frequency contact dimension, and determining the behavior indicator of the contact frequency dimension based on the average communication frequency includes: when the average communication frequency is less than the first set frequency, determining the behavior indicator of the low-frequency contact dimension; when the average communication frequency is greater than or equal to the first set frequency and less than the second set frequency, determining the behavior indicator of the medium-frequency contact dimension; and when the average communication frequency is greater than or equal to the second set frequency, determining the behavior indicator of the high-frequency contact dimension; wherein the first set frequency is less than the second set frequency.

[0188] In one embodiment, the communication behavior dimension also includes a communication dimension in a specific time period and a social circle interaction behavior dimension; when the computer program is executed by the processor to implement the above-mentioned steps of extracting relationship feature indicators of each dimension from multiple relationship data based on the set multiple dimensions, the following steps are specifically implemented: obtaining the communication data of the first person involved in the case and the second person involved in the case in a specific time period from multiple relationship data, and determining the behavior indicators of the communication dimension in the specific time period based on the communication data in the specific time period; obtaining the communication data of the social circle interaction behavior of the first person involved in the case and the second person involved in the case from multiple relationship data, and determining the behavior indicators of the social circle interaction behavior dimension based on the communication data of the social circle interaction behavior.

[0189] In one embodiment, when the computer program is executed by the processor to implement the above-mentioned steps of extracting relationship characteristic indicators of each dimension from multiple relationship data based on the set multiple dimensions, the following steps are specifically implemented: extracting the number of common friends and the number of people belonging to the same community of the first person involved in the case and the second person involved in the case from the multiple relationship data; determining the relationship characteristic indicators of the friend dimension of the people involved in the case based on the number of common friends and the number of people belonging to the same community.

[0190] In one embodiment, when the computer program is executed by the processor to implement the above-mentioned steps of extracting relationship feature indicators of each dimension from multiple relationship data based on the set multiple dimensions, the following steps are specifically implemented: extracting one or more of the voice call data, video call data, mobile phone call data, call frequency data and call duration data of the first person involved in the case and the second person involved in the case from the multiple relationship data; determining the relationship feature indicators of the call behavior dimension based on one or more of the voice call data, video call data, mobile phone call data, call frequency data and call duration data.

[0191] In one embodiment, when the computer program is executed by the processor to implement the above-mentioned steps of extracting relationship characteristic indicators of each dimension from multiple relationship data based on the set multiple dimensions, the following steps are specifically implemented: extracting one or more of the frequency of fund transactions, number of days for fund transactions, fund transaction activity, fixed fund transaction data, fund transaction regularity data, and large-value resource transaction data of the first person involved in the case and the second person involved in the case from the multiple relationship data; determining the relationship characteristic indicators of the fund transaction behavior dimension based on one or more of the frequency of fund transactions, number of days for fund transactions, fund transaction activity, fixed fund transaction data, fund transaction regularity data, and large-value resource transaction data.

[0192] Those skilled in the art will appreciate that all or part of the processes in the above-mentioned embodiments can be implemented by instructing the relevant hardware through a computer program. The computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).

[0193] The technical features of the above embodiments can be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0194] The above-described embodiments merely represent several implementation methods of the present application. While the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the present invention. It should be noted that a person skilled in the art could make various modifications and improvements without departing from the spirit of the present application, all of which fall within the scope of protection of the present application. Therefore, the scope of protection of the present patent application shall be determined by the appended claims.

Claims

1. A relationship closeness analysis method, characterized in that: The method comprises: Acquire multiple relationship data between the first person involved in the case and the second person involved in the case; Extracting relationship characteristic indicators of each dimension from the multiple relationship data based on the set multiple dimensions, the multiple dimensions include a communication behavior dimension, a friend dimension of persons involved in the case, a call behavior dimension, a fund transaction behavior dimension, and a location sending behavior dimension, and the relationship characteristic indicators of each dimension include a communication behavior indicator, a friend characteristic indicator, a voice call behavior indicator, a fund transaction behavior indicator, and a location sending behavior indicator; The relationship characteristic indicators of each dimension are input into the trained classifier model to obtain the relationship closeness output by the trained classifier model, wherein the trained classifier model is obtained by model training through sample relationship characteristic indicators and the relationship closeness between persons involved in the case corresponding to the sample relationship characteristic indicators.

2. The method according to claim 1, characterized in that The communication behavior dimension includes a communication activity dimension, a contact density dimension, and a contact frequency dimension; The extracting the relationship feature index of each dimension from the plurality of relationship data based on the set plurality of dimensions includes: Acquire the communication time between the first person involved in the case and the second person involved in the case from the plurality of relationship data, and determine the behavior indicator of the communication activity dimension according to the collection time of the plurality of relationship data and the communication time; Obtaining the number of days between the first contact and the last contact between the first person involved in the case and the second person involved in the case from the plurality of relationship data, and determining the behavior indicator of the contact density dimension according to the number of days between the first person involved in the case and the last contact; The average communication frequency between the first person involved in the case and the second person involved in the case is obtained from a plurality of relationship data, and the behavior indicator of the contact frequency dimension is determined according to the average communication frequency.

3. The method according to claim 2, characterized in that The communication activity dimension includes a recent communication activity dimension, a mid-term communication activity dimension, and a long-term communication activity dimension. The behavior indicator of the communication activity dimension is determined according to the collection time of the plurality of relationship data and the communication time, including: When the difference between the collection time and the communication time is less than or equal to a first time threshold, the behavior index of the recent communication activity dimension is determined; when the difference between the collection time and the communication time is greater than the first time threshold and less than or equal to a second time threshold, the behavior index of the medium-term communication activity dimension is determined; when the difference between the collection time and the communication time is greater than the second time threshold, the behavior index of the long-term communication activity dimension is determined; and / or, The contact density dimension includes an occasional contact dimension, a regular contact dimension, and a close contact dimension. The behavioral indicator for determining the contact density dimension according to the interval days includes: When the number of communication days in the interval days is less than the first set number of days, or the number of communication days in the interval days is less than the product of the time interval days and the first set value, the behavior index of the occasional communication dimension is determined; when the number of communication days in the interval days is less than the second set number of days, and the number of communication days in the interval days is greater than the product of the time interval days and the first set value and less than the product of the time interval days and the second set value, the behavior index of the regular communication dimension is determined; when the number of communication days in the interval days is less than the second set number of days, and the number of communication days in the interval days is greater than the product of the time interval days and the second set value, the behavior index of the close communication dimension is determined; Wherein, the first setting value is greater than 0 and less than 1, the second setting value is greater than the first setting value and less than 1, and the second setting number of days is greater than the first setting number of days; and / or, The contact frequency dimension includes a low-frequency contact dimension, a medium-frequency contact dimension, and a high-frequency contact dimension. The behavior indicator of the contact frequency dimension is determined according to the average communication frequency, including: When the average communication frequency is less than the first set frequency, the behavior index of the low-frequency contact dimension is determined; when the average communication frequency is greater than or equal to the first set frequency and less than the second set frequency, the behavior index of the medium-frequency contact dimension is determined; when the average communication frequency is greater than or equal to the second set frequency, the behavior index of the high-frequency contact dimension is determined; The first set frequency is smaller than the second set frequency.

4. The method according to claim 2, characterized in that: The communication behavior dimension also includes a communication dimension during a specific period of time and a social circle interaction behavior dimension; The extracting the relationship feature index of each dimension from the plurality of relationship data based on the set plurality of dimensions includes: Acquire communication data of the first person involved in the case and the second person involved in the case during a specific time period from a plurality of relationship data, and determine a behavior indicator of the communication dimension during the specific time period according to the communication data during the specific time period; The communication data of the social circle interaction behaviors of the first person involved in the case and the second person involved in the case are obtained from a plurality of relationship data, and the behavior indicators of the social circle interaction behavior dimensions are determined according to the communication data of the social circle interaction behaviors.

5. The method according to claim 1, characterized in that The extracting the relationship feature index of each dimension from the plurality of relationship data based on the set plurality of dimensions includes: Extracting the number of common friends and the number of friends belonging to the same community of the first person involved in the case and the second person involved in the case from the plurality of relationship data; The relationship characteristic index of the friend dimension of the person involved in the case is determined according to the number of common friends and the number of people belonging to the same community.

6. The method according to claim 1, characterized in that The extracting the relationship feature index of each dimension from the plurality of relationship data based on the set plurality of dimensions includes: Extracting one or more of the voice call data, video call data, mobile phone call data, call frequency data, and call duration data of the first person involved in the case and the second person involved in the case from the plurality of relationship data; The relationship characteristic index of the call behavior dimension is determined based on one or more of the voice call data, video call data, mobile phone call data, call frequency data and call duration data.

7. The method according to claim 1, characterized in that The extracting the relationship feature index of each dimension from the plurality of relationship data based on the set plurality of dimensions includes: Extracting one or more of the frequency of fund transactions, fund transaction days, fund transaction activity, fixed fund transaction data, fund transaction regularity data, and large-amount resource transaction data between the first person involved in the case and the second person involved in the case from the plurality of relationship data; The relationship characteristic index of the fund transaction behavior dimension is determined based on one or more of the fund transaction frequency, fund transaction days, fund transaction activity, fixed fund transaction data, fund transaction regularity data and large-value resource transaction data.

8. A relationship closeness analysis device, characterized in that: The device comprises: An acquisition module, used to acquire multiple relationship data between the first person involved in the case and the second person involved in the case; An extraction module, configured to extract relationship characteristic indicators of each dimension from the plurality of relationship data based on a plurality of set dimensions, wherein the plurality of dimensions include a communication behavior dimension, a friend dimension of persons involved in the case, a call behavior dimension, a fund transaction behavior dimension, and a location sending behavior dimension, and the relationship characteristic indicators of each dimension include a communication behavior indicator, a friend characteristic indicator, a voice call behavior indicator, a fund transaction behavior indicator, and a location sending behavior indicator; The analysis module is used to input the relationship feature indicators of each dimension into the trained classifier model to obtain the relationship closeness output by the trained classifier model, wherein the trained classifier model is obtained by model training through sample relationship feature indicators and the relationship closeness between the persons involved in the case corresponding to the sample relationship feature indicators.

9. A computer device comprising a memory, a processor and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 7 are implemented.

Images