Cross-platform instant messaging user association method and system based on spatiotemporal trajectory

Abstract

The present invention belongs to the technical field of big data processing, and particularly relates to a method and system for associating users of cross-platform instant messaging based on space-time trajectory. Firstly, the space-time points of each user in a data set that change with time are extracted to obtain the travel trajectory of the user using the platform within the time-varying time period; the user's activity area is determined according to the user's travel trajectory, and the activity area is grid-divided according to a preset network edge length, and the user's travel trajectory time interval is divided into time periods according to a preset time interval, and the space-time points in the user's trajectory are represented by grids to form a user trajectory grid that combines the user's geographic location grid with the time interval; the weight of the user trajectory grid in the user trajectory is calculated based on a TF-IWF model, and the TF-IWF value of each grid is integrated to obtain the user trajectory vector; the user trajectory is associated and matched by the similarity between the trajectory vectors of the users to be associated and based on the nearest neighbor distance, so as to improve the accuracy and reliability of cross-platform user association.

Description

Technical Field

[0001] The present invention belongs to the technical field of big data processing, and in particular relates to a method and system for associating users of cross-platform instant messaging based on spatiotemporal trajectories. Background Art

[0002] With the rapid development of mobile Internet technology and wireless positioning technology, the rapid popularization of mobile smart devices, mobile social networks have gradually emerged and profoundly changed people's lifestyles. The combination of mobile smart devices and traditional social platforms has become a new development trend, and various location-based social networks (LBSNs) have emerged, such as Twitter, WeChat, Momo, Foursquare, Weibo, etc. LBSN can use the location information of smart mobile devices to combine the virtual network world with the real world, providing people with more convenient social channels and personalized services. Location-based social discovery (LBSD) in instant messaging applications allows ordinary users to obtain the nicknames, signatures, genders and other information of other nearby users based on their current location, and obtain the distance information of nearby users. Typical examples of such services include WeChat "Nearby People", Weibo "Nearby", Momo "Nearby People", etc. (collectively referred to as "Nearby People" services). Due to the different functional characteristics of different instant messaging tools, there are certain differences in user behavior data. Matching and identifying the corresponding virtual accounts of the same user in multiple instant messaging tools can fully integrate multi-platform data sets and build complete user information. Research on cross-platform instant messaging user association technology is of great significance for locating malicious instant messaging users.

[0003] Research on cross-platform instant messaging user association can be divided into two categories: based on user attribute information and based on user behavior. At present, the association method based on user attribute information mainly associates users through user attribute information, such as user name, mobile phone number and other information, but this method cannot locate users with inconsistent user information on multiple platforms and users who have not opened the address book permission. Therefore, it is necessary to use other identity characteristics of users for association, such as user behavior information. Existing user association methods based on trajectory features use frequency or co-occurrence methods to calculate the similarity between trajectories to determine whether there is a relationship between users. Although such methods can achieve certain results, due to the lack of consideration of the sequential characteristics of user trajectory locations, the correlation between various geographical locations is ignored, making it difficult to characterize the temporal characteristics of user trajectories. Existing methods lack modeling of trajectory sequence information and do not fully explore the potential laws of the order in which users visit geographical locations, resulting in low algorithm accuracy. Summary of the invention

[0004] To this end, the present invention provides a method and system for associating users in cross-platform instant messaging based on spatiotemporal trajectories, which utilizes user distance information and time information to obtain user trajectories, and associates platform users by processing trajectories and performing two trajectory association matches, thereby improving the accuracy and reliability of cross-platform instant messaging user association.

[0005] According to the design scheme provided by the present invention, a method for associating users of cross-platform instant messaging based on spatiotemporal trajectories includes the following contents:

[0006] For the instant messaging user dataset across social platforms, extract the time-space points of each user that change over time to obtain the user's travel trajectory using the instant messaging social platform during the time period;

[0007] Determine the user's activity area based on the user's travel trajectory, divide the activity area into grids based on the preset network edge length, divide the user's travel trajectory time interval into time periods based on the preset time interval, and represent the time and space points in the user's travel trajectory with grids to form a user trajectory grid that combines the user's geographic location grid with the time interval;

[0008] The weight of the user trajectory grid in the user trajectory is calculated based on the TF-IWF model, and the TF-IWF value of each grid is combined to obtain the user trajectory vector;

[0009] The similarity between the trajectory vectors of the users to be associated is used to preliminarily determine whether two users are associated. Based on the preliminary judgment results, the user trajectories are secondary matched and associated based on the nearest point distance, and the matched users on different social platforms are associated to improve the user trajectory information.

[0010] As a cross-platform instant messaging user association method based on spatiotemporal trajectory of the present invention, further, the user activity area is determined according to the latitude and longitude coordinates of the geographical location where the user's travel trajectory appears, the travel trajectory is grid-divided using the positioning error as the preset grid side length, and the geographical location coordinates of all trajectories in the user's travel trajectory are converted into grid numbers. The trajectory formed by the user trajectory grid is represented as in, represents the user's time-space point, g1,g2,...,g i Indicates the geographic location grid number in the trajectory, i indicates the number of divided grids, t1, t2, ..., t n Indicates the time point when the user appears in the corresponding grid, and n indicates the number of divided time periods.

[0011] As a cross-platform instant messaging user association method based on spatiotemporal trajectory of the present invention, further, the grid number is converted from the geographical location coordinates, and the conversion process is expressed as: i=(r*(g x -1)+g y ), where the latitude range corresponding to the user activity area is represented by (lat1, lat2), the longitude range is represented by (lon1, lon2), the number of grid rows is l, the number of columns is r, and g x Indicates the row number of the grid, g y Indicates the column number where the grid is located, and

[0012] As a cross-platform instant messaging user association method based on spatiotemporal trajectories of the present invention, further, in calculating weights based on the TF-IWF model, firstly, by calculating the frequency value TF of each grid in the trajectory and the co-occurrence frequency value IWF of the grid in all user trajectories to be associated, the frequency value TF and the co-occurrence frequency value IWF of each grid in the trajectory are used as grid weights, and the trajectory vector is constructed by the grid weights.

[0013] As a cross-platform instant messaging user association method based on spatiotemporal trajectory of the present invention, further, the cosine similarity between the trajectory vectors of the users to be associated is used. If the cosine similarity is greater than a preset association threshold, it is preliminarily determined that the users to be associated have an association relationship.

[0014] As a cross-platform instant messaging user association method based on spatiotemporal trajectories of the present invention, further, a cosine similarity calculation formula for trajectory vectors of users i and j to be associated is expressed as: sim(Ti,Tj)=cos(v*(Ti),v*(Tj)), wherein Ti, Tj represent user trajectories corresponding to i and j, respectively, and v*(Ti), v*(Tj) represent trajectory vectors corresponding to user trajectories Ti, Tj, respectively.

[0015] As a cross-platform instant messaging user association method based on spatiotemporal trajectories of the present invention, further, for users preliminarily determined to have an associated relationship, a secondary matching association is performed on the user trajectories based on the nearest neighbor point distance, which includes the following contents: first, the distance between the position point in the user trajectory and the position point in the trajectory of another associated user in the same time period is calculated; then, the most adjacent position point group in the associated user trajectory in the same time period is determined according to the position point distance, and the user trajectory distance set is obtained according to the most adjacent position point group in each time period, and the user trajectory is secondary matched through the user trajectory distance set.

[0016] Furthermore, the present invention also provides a cross-platform instant messaging user association system based on spatiotemporal trajectory, comprising: a data processing module, a grid division module, a vector acquisition module and an association determination module, wherein:

[0017] A data processing module is used to extract the time-space points of each user that change over time for the instant messaging user data set across social platforms to obtain the user's travel trajectory using the instant messaging social platform during the time period of change;

[0018] A grid division module is used to determine the user's activity area according to the user's travel trajectory, and to divide the activity area into grids according to the preset network edge length, divide the user's travel trajectory time interval into time periods according to the preset time interval, and represent the time and space points in the user's travel trajectory with grids to form a user trajectory grid that combines the user's geographical location grid with the time interval;

[0019] A vector acquisition module is used to calculate the weight of the user trajectory grid in the user trajectory based on the TF-IWF model, and to obtain the user trajectory vector by integrating the TF-IWF values ​​of each grid;

[0020] The association determination module is used to preliminarily determine whether two users are associated based on the similarity between the trajectory vectors of the users to be associated; and based on the preliminary judgment result, the user trajectories are secondary matched and associated based on the nearest point distance, and the matched users on different social platforms are associated to improve the user trajectory information.

[0021] Beneficial effects of the present invention:

[0022] The present invention represents the user's trajectory in the form of a vector, and preliminarily determines whether the user trajectories of different platforms match by calculating the cosine similarity between vectors; and performs secondary trajectory distance matching based on the distance between matching points in the trajectory calculated based on the nearest neighbor point distance. By matching the user's trajectory information in multiple different instant messaging tools, the accuracy and reliability of user association are improved, malicious instant messaging users are screened and filtered, and the application in actual scenarios is facilitated. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] Figure 1 This is a schematic diagram of the process of a method for associating users of cross-platform instant messaging based on spatiotemporal trajectories in an embodiment;

[0024] Figure 2 This is an illustration of secondary matching of user trajectories in the embodiment. DETAILED DESCRIPTION

[0025] In order to make the purpose, technical solutions and advantages of the present invention clearer and more understandable, the present invention is further described in detail below in conjunction with the accompanying drawings and technical solutions.

[0026] When users use the "nearby people" function in LBSNs, LBSNs will require users to grant the application permission to obtain geographic location. LBSN applications obtain the user's precise geographic location and upload it to the server. The server informs the user of the precise distance between the user and other users in the form of relative distance after blurring. In this way, users can only obtain the blurred relative distance between themselves and other users, but cannot obtain the precise distance. With the application of LBSD user location information privacy protection technology, LBSD services gradually use distance segments to notify the distance between users. Cross-platform user association aims to identify the real user identity behind the account in the absence of a unified user identity, and associate accounts belonging to the same real user on multiple platforms, providing a feasible solution for integrating multi-platform data and expanding the scope of instant communication positioning users. At present, the research on cross-platform instant communication user association can be divided into two categories: based on user attribute information and based on user behavior. The method based on user attribute information mainly uses multiple attribute information such as user name for identity recognition and multi-platform user information association based on mobile phone number, but when the multi-platform user information is inconsistent and the user does not open the address book permission, user association cannot be performed. The method based on user behavior mainly calculates the similarity of user trajectories through features such as geographic location frequency and co-occurrence number, but this method lacks the use of time information, resulting in the inability to fully explore the sequential characteristics of the trajectory, affecting the accuracy of trajectory matching. In order to solve the problem of multi-platform user association based on mobile phone numbers without opening the address book permission, and at the same time make full use of the time information in the user trajectory, the embodiment of the present invention provides a cross-platform instant messaging user association method based on spatiotemporal trajectory, see Figure 1 As shown, it contains the following contents:

[0027] S101. For a cross-social platform instant messaging user dataset, extract the time-space points of each user that change over time to obtain the user's travel trajectory using the instant messaging social platform within the time-varying period;

[0028] S102, determining the user's activity area according to the user's travel trajectory, and dividing the activity area into grids according to a preset network edge length, dividing the user's travel trajectory time interval into time periods according to a preset time interval, and representing the time and space points in the user's travel trajectory with grids to form a user trajectory grid that combines the user's geographical location grid with the time interval;

[0029] S103, using the TF-IWF model to calculate the weight of the user trajectory grid in the user trajectory, and synthesizing the TF-IWF values ​​of each grid to obtain the user trajectory vector;

[0030] S104, preliminarily judging whether two users are associated by the similarity between the trajectory vectors of the users to be associated; and performing secondary matching association on the user trajectories based on the nearest point distance according to the preliminary judgment result, and associating the matched and associated users of different social platforms to improve the user trajectory information.

[0031] Cross-platform user association based on user attribute features mainly associates users through multiple information such as user attribute information such as mobile phone number, user name, home address, Tel, Email, personal signature, etc. User attributes such as user name, user avatar, and geographic location are used for identity recognition, and supervised classification models such as decision tree and SVM are used for training to associate cross-platform users; or, the Jaro distance and Wordnet ontology dictionary are used to calculate the similarity of attribute values, and the user matching on Twitter and Linkedln is realized by training the naive Bayes classifier; or, by analyzing the internal structure of instant messaging tools, the user information of various instant messaging tools is accurately and standardizedly crawled based on the precise search of graphical user interface elements; then, based on the mobile phone number search and friend recommendation functions of instant messaging tools, the user information corresponding to the mobile phone number is obtained, and users on different platforms are associated; then, the user information corresponding to the mobile phone number is compared with the user information of "nearby people" to achieve the association between "nearby people" users and mobile phone numbers; finally, the distance information reported by the user is obtained and combined with a high-precision positioning algorithm for positioning, and then the user information of multiple instant messaging tools is integrated. Cross-platform user association based on user behavior characteristics mainly associates users through user behavior characteristics such as activity trajectory, friend relationship and other information, and measures the degree of connection between the trajectory and the query location through the similarity function. At the same time, the spatial distance and order constraints are considered. The Incrementalk-NN based Algorithm (IKNN) is analyzed to effectively prune and refine the trajectory by using the designed similarity lower and upper limits, and the feasibility of using a general spatial index based on a simple incremental k-NN algorithm to achieve efficient k-BCT search is analyzed. The user trajectory information partially reflects the user's movement trajectory in the real world, which becomes the behavioral information with strong characterization of the user's identity. The entire map can be divided into many grids, and then each user's trajectory is represented as a sequence of several small grids. The Term Frequency-Inverse Document Frequency (TF-IDF) model is used to convert each user's trajectory into a vector, and the similarity between user trajectories is obtained by calculating the cosine similarity between vectors. The PV-DM model in the Paragraph2vec algorithm can be used to extract the location access order features in the trajectory sequence, obtain the vector representation of the user trajectory, and finally determine whether the trajectory matches through the user trajectory vector; or, based on the kernel density estimation (KDE) method, the accuracy and efficiency can be improved by improving the data sparsity problem in the user similarity measurement. The location data is organized based on the grid structure to prune the search space. In cross-platform user association, when the user's attribute information is used for association, the user's attribute information will change over time, and too large a data set will affect the association results.At the same time, when using the user's trajectory information for association, the correlation between the various geographical locations is ignored, and the sequential characteristics of the various geographical locations in the trajectory are not fully explored. In the embodiment of this case, the sequential access characteristics and distance characteristics of the user's trajectory are considered, which can solve the problem of insufficient utilization of geographical location frequency and time information, increase the number, accuracy and reliability of users that can be located in cross-platform instant messaging, and improve the effectiveness of user account association.

[0032] Because of the errors in the positioning algorithm, when users use instant messaging tools at the same location, the errors will cause inconsistencies in the positioning results expressed in longitude and latitude, resulting in sparse samples and affecting the matching results. Therefore, in the embodiment of this case, the area within the geographical range where the user appears is divided into a number of small grids with appropriate accuracy, and the user positioning results are replaced by small grids. This not only avoids the problem of sparse samples, but also more intuitively reflects the user's geographical location, making it easier to calculate the similarity of the trajectory. Furthermore, in the embodiment of this case, the user's activity area is determined based on the longitude and latitude coordinates of the geographical location where the user's travel trajectory appears, the travel trajectory is gridded using the positioning error as the preset grid side length, and the geographical location coordinates of all trajectories in the user's travel trajectory are converted into grid numbers. The trajectory formed by the user trajectory grid is represented as in, represents the user's space-time point, g1,g2,...,g i Indicates the geographic location grid number in the trajectory, i indicates the number of divided grids, t1, t2, ..., t n Indicates the time point when the user appears in the corresponding grid, and n indicates the number of time periods. The time period can be divided according to the actual situation. For example, a day can be divided into 24 time periods at an interval of one hour to form the time point when the user appears on the grid. The grid number is converted from the geographic location coordinates. The conversion process is expressed as: i =(r*(g x -1)+g y ), where the latitude range corresponding to the user activity area is represented by (lat1, lat2), the longitude range is represented by (lon1, lon2), the number of grid rows is l, the number of columns is r, and g x Indicates the row number of the grid, g y Indicates the column number where the grid is located, and

[0033] As a cross-platform instant messaging user association method based on spatiotemporal trajectories in an embodiment of the present invention, further, in calculating the weights based on the TF-IWF model, firstly, by calculating the frequency value TF of each grid in the trajectory and the co-occurrence frequency value IWF of the grid in all user trajectories to be associated, the frequency value TF and the co-occurrence frequency value IWF of each grid in the trajectory are used as grid weights, and the trajectory vector is formed by the grid weights.

[0034] The trajectory vector calculation method based on the TF-IWF model is shown in Algorithm 1 and can be divided into the following three steps:

[0035] Step 1 Calculate the frequency value TF of the i-th grid in trajectory j. TF is defined as:

[0036]

[0037] Among them, n i,j Represents the grid g i In the trajectory T j The denominator represents the number of times it appears in the trajectory T j The sum of the frequencies of all grids in the grid. The TF value can be used to represent the grid g i The weight in the trajectory.

[0038] Step 2 calculates the co-occurrence frequency value IWF of the grid in the two trajectories. IWF is defined as:

[0039]

[0040] The molecular part is represented by the grid g i The denominator is the sum of all grid frequencies in the combined spatiotemporal set after the two trajectories are merged. This weighting method increases the influence of the co-occurring grids in the two trajectories on the trajectory similarity weight, and more accurately expresses the importance of this grid in trajectory matching.

[0041] Step 3 Calculate trajectory T j The TF-IWF value of each grid in . Construct the trajectory vector v(T j ), defined as:

[0042]

[0043] Normalize this vector to

[0044]

[0045] v*(T j ) is used as the user trajectory vector for similarity calculation.

[0046] Table 1: User Trajectory Vector Generation Algorithm Based on TF-IWF

[0047]

[0048] The user trajectory vector is calculated using the TF-IWF model. The weighted method of calculating the frequency twice not only increases the weight of high-frequency grids in the trajectory but also raises the proportion of co-occurring grids in the trajectory group to be compared, more precisely expressing the importance of co-occurring high-frequency grids in trajectory matching. In the embodiments of this case, the user trajectory vector containing the grid sequence is finally obtained through the weight calculation and adjustment of the grids, and the similarity between the user trajectory vectors is calculated to determine whether two users are related. The calculation of vector similarity can be achieved through the cosine similarity of two trajectory vectors:

[0049] sim(Ti,Tj) = cos(v*(Ti),v*(Tj)) (8)

[0050] If the cosine similarity of the user trajectories is greater than the threshold α, it is preliminarily determined that these two users are related.

[0051] As the cross-platform instant messaging user association method based on spatio-temporal trajectories in the embodiments of the present invention, further, for users preliminarily determined to be related, the user trajectories are secondarily matched and associated based on the nearest point distance, including the following: First, calculate the distance between the position points in the user trajectory within the same time period and the position points in the trajectory of another associated user; then, determine the group of the nearest position points in the trajectory of the associated user within the same time period according to the position point distance, and obtain the user trajectory distance set based on the group of the nearest position points in each time period, and perform the secondary matching of the user trajectories through the user trajectory distance set.

[0052] Due to the protection of user location information by instant messaging tools, errors will inevitably occur when performing trilateration positioning and spatial division positioning based on statistical features for users. When there are no co-occurring grids among all the positions of the user trajectory, it is difficult to perform weighted analysis by statistically counting co-occurring grids. Therefore, the user trajectory can be secondarily matched based on the nearest point distance of the user trajectory, as Figure 2 shown. Calculate the distances L1, L2, L3... between the position points in one trajectory and the position points in another trajectory in sequence within the same time period T1, and find the group of the nearest points (L2 < L3 < L1) in the same time period of the two trajectories, that is, the group of points connected by L2. And calculate the distances between each group of points after this group of points to obtain the user trajectory distance set for the secondary matching of the user trajectories. Among them, the content of the secondary matching algorithm of the user trajectory can be designed as follows:

[0053] Step 1: Define the maximum distance between each set of matching points of two matching trajectories to be less than the side length g of the small grid.

[0054] Step 2 Calculate the first group of points with similar distances in the first period of the two trajectories, and record the distance between the two points as d1. Calculate and record the distance between each group of points after the first group of points in the trajectory as d2, d3, d4, etc. This method of matching trajectory distances based on the nearest point emphasizes the distance characteristics between trajectories and the sequential characteristics of user trajectories, thereby increasing the accuracy of user trajectory matching.

[0055] Step3 Get the distance set between two adjacent trajectory points as follows:

[0056] Dis(d1,d2,d3,d4,...)

[0057] Step 4: Statistically calculate the matching degree of the trajectory point distance set and the maximum matching distance g for the secondary matching user trajectory.

[0058] Table 2: Secondary matching algorithm of user trajectory

[0059]

[0060]

[0061] By judging whether users have an association relationship based on the similarity of user trajectory vectors, effective cross-platform association can be achieved. After associating users with matching relationships, the user trajectories obtained from different platforms are integrated to improve the user trajectory information, and the location information of multiple platforms is used to supplement the user's positioning, thereby increasing the scale of locatable users.

[0062] Furthermore, based on the above method, an embodiment of the present invention also provides a cross-platform instant messaging user association system based on spatiotemporal trajectory, comprising: a data processing module, a grid division module, a vector acquisition module and an association determination module, wherein:

[0063] A data processing module is used to extract the time-space points of each user that change over time for the instant messaging user data set across social platforms to obtain the user's travel trajectory using the instant messaging social platform during the time period of change;

[0064] A grid division module is used to determine the user's activity area according to the user's travel trajectory, and to divide the activity area into grids according to the preset network edge length, divide the user's travel trajectory time interval into time periods according to the preset time interval, and represent the time and space points in the user's travel trajectory with grids to form a user trajectory grid that combines the user's geographical location grid with the time interval;

[0065] A vector acquisition module is used to calculate the weight of the user trajectory grid in the user trajectory based on the TF-IWF model, and to obtain the user trajectory vector by integrating the TF-IWF values ​​of each grid;

[0066] The association determination module is used to preliminarily determine whether two users are associated based on the similarity between the trajectory vectors of the users to be associated; and based on the preliminary judgment result, the user trajectories are secondary matched and associated based on the nearest point distance, and the matched users on different social platforms are associated to improve the user trajectory information.

[0067] By matching the trajectories of users of different instant messaging tools, information association of different instant messaging platforms can be achieved. In the process of trajectory matching, not only the similarity between sets is calculated based on frequency or co-occurrence, but also the sequential characteristics of user trajectory positions and the distance characteristics of user trajectory positions are taken into account. When associating different instant messaging tools, the accuracy of locating users can be improved, and the number of locatable users can be increased at the same time, which has good application prospects.

[0068] Unless otherwise specifically stated, the relative steps, numerical expressions and values ​​of the components and steps set forth in these embodiments do not limit the scope of the present invention.

[0069] Based on the above method and / or system, an embodiment of the present invention also provides a server, including: one or more processors; a storage device for storing one or more programs, when the one or more programs are executed by the one or more processors, the one or more processors implement the above method.

[0070] Based on the above method and / or system, an embodiment of the present invention further provides a computer-readable medium having a computer program stored thereon, wherein the above method is implemented when the program is executed by a processor.

[0071] In all examples shown and described herein, any specific values ​​should be interpreted as merely exemplary and not as limiting, and thus other examples of the exemplary embodiments may have different values.

[0072] It should be noted that similar reference numerals and letters denote similar items in the following drawings, and therefore, once an item is defined in one drawing, further definition and explanation thereof is not required in subsequent drawings.

[0073] Finally, it should be noted that the above-described embodiments are only specific implementations of the present invention, which are used to illustrate the technical solutions of the present invention, rather than to limit them. The protection scope of the present invention is not limited thereto. Although the present invention is described in detail with reference to the above-described embodiments, ordinary technicians in the field should understand that any technician familiar with the technical field can still modify the technical solutions recorded in the above-described embodiments within the technical scope disclosed by the present invention, or can easily think of changes, or make equivalent replacements for some of the technical features therein; and these modifications, changes or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present invention, and should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention shall be subject to the protection scope of the claims.

Claims

1. A cross-platform instant messaging user association method based on spatiotemporal trajectory, characterized in that: Contains the following: For the instant messaging user dataset across social platforms, extract the time-space points of each user that change over time to obtain the user's travel trajectory using the instant messaging social platform during the time period; Determine the user's activity area based on the user's travel trajectory, divide the activity area into grids based on the preset network edge length, divide the user's travel trajectory time interval into time periods based on the preset time interval, and represent the time and space points in the user's travel trajectory with grids to form a user trajectory grid that combines the user's geographic location grid with the time interval; The weight of the user trajectory grid in the user trajectory is calculated based on the TF-IWF model, and the TF-IWF value of each grid is combined to obtain the user trajectory vector; wherein, in the weight calculation based on the TF-IWF model, the frequency value TF of each grid in the trajectory and the co-occurrence frequency value IWF of the grid in all the user trajectories to be associated are first calculated, and the frequency value TF and the co-occurrence frequency value IWF of each grid in the trajectory are used as the grid weight, and the trajectory vector is formed by the grid weight; The similarity between the trajectory vectors of the users to be associated is used to preliminarily determine whether two users are associated; and based on the preliminary judgment results, the user trajectories are secondary matched and associated based on the nearest point distance, and the users of different social platforms that are matched and associated are associated to improve the user trajectory information, wherein the secondary matching and association process includes: first, defining the maximum distance of each group of matching points of the two matching trajectories to be less than the grid side length g, and calculating the distance between the position point in the user trajectory and the position point in the trajectory of another associated user in the same time period; then, the most adjacent position point group in the associated user trajectory in the same time period is determined based on the position point distance, and the user trajectory distance set is obtained based on the most adjacent position point group in each time period, and the user trajectory distance set is used to perform secondary matching of the user trajectory, and the user trajectory matching degree of the secondary matching of the user trajectory distance set and the maximum matching distance g is statistically calculated.

2. The method for associating users of cross-platform instant messaging based on spatiotemporal trajectories according to claim 1, characterized in that: The user activity area is determined based on the longitude and latitude coordinates of the geographical location where the user's travel trajectory appears. The travel trajectory is gridded using the positioning error as the preset grid side length. The geographical location coordinates of all trajectories in the user's travel trajectory are converted into grid numbers. The trajectory formed by the user's trajectory grid is represented as in, represents the user's time-space point, g1,g2,...,g i Indicates the geographic location grid number in the trajectory, i indicates the number of divided grids, t1, t2, ..., t n Indicates the time point when the user appears in the corresponding grid, and n indicates the number of divided time periods.

3. The method for associating users of cross-platform instant messaging based on spatiotemporal trajectories according to claim 1 or 2, characterized in that: The grid number is converted from the geographic location coordinates. The conversion process is expressed as: i =(r*(g x -1)+g y ), where the latitude range corresponding to the user activity area is represented by (lat1, lat2), the longitude range is represented by (lon1, lon2), the number of grid rows is l, the number of columns is r, and g x Indicates the row number of the grid, g y Indicates the column number where the grid is located, and 4. The method for associating users of cross-platform instant messaging based on spatiotemporal trajectories according to claim 1, characterized in that: The cosine similarity between the trajectory vectors of the users to be associated is used. If the cosine similarity is greater than a preset association threshold, it is preliminarily determined that the users to be associated have an associated relationship.

5. The method for associating users of cross-platform instant messaging based on spatiotemporal trajectories according to claim 4 is characterized in that: The formula for calculating the cosine similarity of the trajectory vectors of users i and j to be associated is expressed as: sim(Ti,Tj)=cos(v*(Ti),v*(Tj)), where Ti,Tj represent the user trajectories corresponding to i and j respectively, and v*(Ti) and v*(Tj) represent the trajectory vectors corresponding to user trajectories Ti,Tj respectively.

6. A cross-platform instant messaging user association system based on spatiotemporal trajectory, characterized in that: It includes: data processing module, grid division module, vector acquisition module and association determination module, among which, A data processing module is used to extract the time-space points of each user that change over time for the instant messaging user data set across social platforms to obtain the user's travel trajectory using the instant messaging social platform during the time period of change; A grid division module is used to determine the user's activity area according to the user's travel trajectory, and to divide the activity area into grids according to the preset network edge length, divide the user's travel trajectory time interval into time periods according to the preset time interval, and represent the time and space points in the user's travel trajectory with grids to form a user trajectory grid that combines the user's geographical location grid with the time interval; The vector acquisition module is used to calculate the weight of the user trajectory grid in the user trajectory based on the TF-IWF model, and to obtain the user trajectory vector by integrating the TF-IWF values ​​of each grid; wherein, in calculating the weight based on the TF-IWF model, firstly, by calculating the frequency value TF of each grid in the trajectory and the co-occurrence frequency value IWF of the grid in all the user trajectories to be associated, the frequency value TF and the co-occurrence frequency value IWF of each grid in the trajectory are used as the grid weight, and the trajectory vector is formed by the grid weight; The association determination module is used to preliminarily determine whether two users are associated through the similarity between the trajectory vectors of the users to be associated; and for the preliminary judgment result, the user trajectories are secondary matched and associated based on the nearest point distance, and the users of different social platforms that are matched and associated are associated to improve the user trajectory information, wherein the secondary matching and association process includes: first, defining the maximum distance of each group of matching points of the two matching trajectories to be less than the grid side length g, and calculating the distance between the position point in the user trajectory and the position point in the trajectory of another associated user in the same time period; then, according to the position point distance, the most adjacent position point group in the associated user trajectory in the same time period is determined, and the user trajectory distance set is obtained according to the most adjacent position point group in each time period, and the user trajectory distance set is used to perform secondary matching of the user trajectory, and the user trajectory matching degree of the secondary matching of the user trajectory distance set and the maximum matching distance g is statistically calculated.

7. 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 method according to any one of claims 1 to 5 is implemented.

8. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the method according to any one of claims 1 to 5 is implemented.

Images