User occupational recognition method and device, equipment, storage medium and program product

Abstract

The invention discloses a user occupational recognition method and device, equipment, a storage medium and a program product, relates to the technical field of computers, and discloses a user occupational recognition method which comprises the following steps: generating a personal trajectory vector set corresponding to a user terminal according to a dynamic link relationship between the user terminal and a base station, generating a time code and a position code of the personal trajectory vector set; pre-training the base large model based on the time code and the position code, and extracting static features of a user and static features of a space contained in the pre-trained base large model; according to the static features of the user and the static features of the space, training a pre-constructed occupational classifier to obtain a trained occupational classifier; the trained occupational classifier is used for user occupational recognition. According to the invention, the user occupational recognition accuracy is improved.

Description

Technical Field

[0001] The present application relates to the field of computer technology, and in particular to a method, apparatus, device, storage medium and program product for identifying a user's occupation. Background Art

[0002] In traditional solutions, the means of determining user occupations based on data include: building a classification model to predict occupations based on the application categories used by users, application installation records, and application usage data. The above methods are difficult to collect data, complex to construct feature vectors, and take a long time, resulting in low accuracy in the final prediction of user occupations. Summary of the invention

[0003] The main purpose of this application is to provide a method, device, equipment, storage medium and program product for identifying a user's occupation, aiming to solve the technical problem of low accuracy in predicting the user's occupation.

[0004] To achieve the above objectives, the present application proposes a method for identifying a user's occupation, the method comprising:

[0005] Generate a personal trajectory vector set corresponding to the user terminal according to the dynamic link relationship between the user terminal and the base station, and generate a time code and a position code of the personal trajectory vector set;

[0006] Pre-training the base large model based on the time code and the position code, and extracting static features of the user and static features of the space contained in the pre-trained base large model;

[0007] According to the static features of the user and the static features of the space, a pre-constructed occupation classifier is trained to obtain a trained occupation classifier; the trained occupation classifier is used to identify the user's occupation.

[0008] In one embodiment, the step of generating a personal trajectory vector set corresponding to the user terminal according to the dynamic link relationship between the user terminal and the base station includes:

[0009] Generate a ternary grid group of the user terminal according to the dynamic link relationship between the user terminal and the base station, the ternary grid group including the terminal number, the grid identifier of the grid area where the base station is located, and the time when the user terminal is located at the grid identifier;

[0010] A set of personal trajectory vectors with timestamps is generated according to the triple grid group.

[0011] In one embodiment, before the step of generating a ternary mesh group of the user terminal according to the dynamic link relationship between the user terminal and the base station, the step further includes:

[0012] Get the target area;

[0013] The target area is divided into grids to form a grid set covering the target area, wherein the grid set includes at least one grid area, and each of the grid areas corresponds to a grid identifier.

[0014] In one embodiment, the step of generating a ternary mesh group of the user terminal according to the dynamic link relationship between the user terminal and the base station includes:

[0015] Generate the terminal number, the longitude and latitude of the base station and the time when the user terminal is at the longitude and latitude of the base station according to the dynamic link relationship between the user terminal and the base station;

[0016] Determine the grid identifier where the base station is located according to the latitude and longitude of the base station;

[0017] A ternary grid group of the user terminal is generated according to the grid identifier where the base station is located, the terminal number and the time.

[0018] In one embodiment, the step of training a pre-built occupation classifier according to the static features of the user and the static features of the space includes:

[0019] Obtaining a true value sample set of occupations, and constructing a training trajectory vector set according to the true value sample set of occupations;

[0020] Generating a time code and a position code of the training trajectory vector set;

[0021] A pre-built occupation classifier is trained based on the time coding and position coding of the training trajectory vector set, as well as the static features of the user and the static features of the space.

[0022] In one embodiment, the step of pre-training the base large model based on the time code and the position code comprises:

[0023] Constructing a pre-training task, the pre-training task comprising predicting missing grid identifiers and corresponding link times in the personal trajectory vector set, the missing grid identifiers being dynamically masked grid identifiers of a set proportion;

[0024] Based on the pre-training task, the time code and the position code are input into the base large model for pre-training.

[0025] In addition, to achieve the above purpose, the present application also proposes a user occupation identification device, the user occupation identification device comprising:

[0026] A generating module, configured to generate a personal trajectory vector set corresponding to the user terminal according to a dynamic link relationship between the user terminal and the base station, and to generate a time code and a position code of the personal trajectory vector set;

[0027] An extraction module, used for pre-training the base large model based on the time code and the position code, and extracting the static features of the user and the static features of the space contained in the pre-trained base large model;

[0028] The training module is used to train a pre-built occupation classifier according to the static features of the user and the static features of the space to obtain a trained occupation classifier; the trained occupation classifier is used to identify the user's occupation.

[0029] In addition, to achieve the above-mentioned purpose, the present application also proposes a user occupation identification device, which includes: a memory, a processor, and a computer program stored in the memory and executable on the processor, and the computer program is configured to implement the steps of the user occupation identification method described above.

[0030] In addition, to achieve the above-mentioned purpose, the present application also proposes a storage medium, which is a computer-readable storage medium, and a computer program is stored on the storage medium. When the computer program is executed by a processor, the steps of the user occupation identification method described above are implemented.

[0031] In addition, to achieve the above-mentioned purpose, the present application also provides a computer program product, which includes a computer program, and when the computer program is executed by a processor, the steps of the user occupation identification method described above are implemented.

[0032] One or more technical solutions proposed in this application have at least the following technical effects:

[0033] According to the dynamic link relationship between the user terminal and the base station, a personal trajectory vector set corresponding to the user terminal is generated, and the time code and position code of the personal trajectory vector set are generated; the base large model is pre-trained based on the time code and position code, and the static features of the user and the static features of the space contained in the pre-trained base large model are extracted; according to the static features of the user and the static features of the space, the pre-built occupation classifier is trained to obtain a trained occupation classifier; the trained occupation classifier is used for user occupation identification. The personal trajectory vector set is generated through the dynamic link relationship to reduce the calculation load of the large model, improve the comprehensiveness of the personal trajectory determination, and facilitate the subsequent user occupation identification. The base large model is pre-trained according to the time code and position code to obtain static features to construct a classifier, and the user's occupation is identified by the classifier constructed by the static features shown in the dynamic link relationship between the user and the base station, thereby improving the accuracy of user occupation identification. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present application and, together with the description, serve to explain the principles of the present application.

[0035] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, for ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative labor.

[0036] Figure 1 A flowchart of the first embodiment of the method for identifying occupation of a user in this application;

[0037] Figure 2 A flowchart of the second embodiment of the method for identifying the occupation of a user in this application is provided;

[0038] Figure 3 This is a schematic diagram of the module structure of the user occupation identification device according to an embodiment of the present application;

[0039] Figure 4 This is a schematic diagram of the device structure of the hardware operating environment involved in the user occupation identification method in the embodiment of the present application.

[0040] The purpose, features and advantages of this application will be further described in conjunction with the embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0041] It should be understood that the specific embodiments described herein are only used to explain the technical solutions of the present application and are not used to limit the present application.

[0042] In order to better understand the technical solution of the present application, a detailed description will be given below in conjunction with the accompanying drawings and specific implementation methods.

[0043] The main solution of the embodiment of the present application is: based on the dynamic link relationship between the user terminal and the base station, a personal trajectory vector set corresponding to the user terminal is generated, and the time code and position code of the personal trajectory vector set are generated; based on the time code and position code, the base large model is pre-trained, and the static features of the user and the static features of the space contained in the pre-trained base large model are extracted; according to the static features of the user and the static features of the space, a pre-built occupation classifier is trained to obtain a trained occupation classifier; the trained occupation classifier is used to identify the user's occupation.

[0044] In this embodiment, for the convenience of description, the following description is made with the user occupation identification device as the execution subject.

[0045] The existing methods of determining user occupations based on data include: building a classification model to predict occupations based on the application categories used by users, application installation records, and application usage data. The above methods are difficult to collect data, complex to construct feature vectors, and take a long time, resulting in low accuracy in the final prediction of user occupations.

[0046] The present application provides a solution, which generates a personal trajectory vector set through a dynamic link relationship to reduce the computing load of a large model, improves the comprehensiveness of personal trajectory determination, and facilitates subsequent user occupation identification. The base large model is pre-trained according to time coding and position coding to obtain static features to construct a classifier. The classifier constructed by the static features shown in the dynamic link relationship between the user and the base station is used to identify the user's occupation, thereby improving the accuracy of user occupation identification.

[0047] It should be noted that the execution subject of this embodiment may be a computing service device with data processing, network communication and program running functions, such as a tablet computer, a personal computer, a mobile phone, etc., or an electronic device capable of realizing the above functions, a user occupation identification device, etc. The following takes the user occupation identification device as an example to illustrate this embodiment and the following embodiments.

[0048] Based on this, the present application embodiment provides a method for identifying a user's occupation. Figure 1 , Figure 1 This is a flow chart of the first embodiment of the user occupation identification method of the present application.

[0049] In this embodiment, the user occupation identification method includes steps S10 to S30:

[0050] Step S10: generating a personal trajectory vector set corresponding to the user terminal according to the dynamic link relationship between the user terminal and the base station, and generating a time code and a position code of the personal trajectory vector set.

[0051] This embodiment can be applied to planning and implementation fields such as smart cities, advertising marketing, infrastructure and commercial supporting facilities. In this embodiment, the dynamic link relationship between the user terminal and the base station obtained is information obtained by the user after the terminal is authorized.

[0052] It should be noted that the dynamic link relationship means that the location and signal strength of the user terminal change over time, and the connection between the user terminal and the base station is dynamic. The dynamic link relationship includes parameters such as the mobile phone number, the longitude and latitude of the base station, and the link time of the user terminal at the current longitude and latitude. The personal trajectory vector set is a vector set corresponding to the user terminal. The vectors in the set describe the user's spatial location, activity mode, movement path and other characteristics.

[0053] Optionally, the signal strength, location data, and base station information in the dynamic connection relationship are obtained, and the dynamic link relationship is processed, including feature extraction, location mapping, and vector construction, to generate a personal trajectory vector set corresponding to the user terminal. Optionally, the vector in the personal trajectory vector set may include the user's current location, speed, moving direction, time information, etc.

[0054] It should be noted that the signal strength is the signal strength between the user terminal and multiple base stations. The signal strength is used to estimate the distance and build the user's movement trajectory in combination with the timestamp. The location data is the actual location of the user terminal obtained in combination with GPS (Global Positioning System) data or other location tracking technologies. Base station information includes the location information and coverage of the base station.

[0055] Feature extraction is to extract features from signal strength and location data, such as movement speed, dwell time, type of base station visited, etc. Location mapping is to map location data into a high-dimensional space, taking into account the movement and dwelling of users between different base stations. Vector construction is to construct the user's spatial vector based on the extracted features.

[0056] Personal trajectory vector sets can also be generated through time series modeling, clustering and classification, and personalized features. Among them, time series modeling uses time series data analysis techniques such as time series neural networks and recursive neural networks to model the user's mobility pattern and generate spatial vectors of time series. Clustering and classification is to perform cluster analysis on the generated vectors to classify the user's mobility pattern into different categories. The vectors can also be classified to identify the user's common activity areas or behavior patterns. Personalized features are to generate personalized spatial vector sets based on the user's historical mobility data and behavior. For example, a user may often go to a specific place within a specific time period, and this information can be integrated into the vector.

[0057] Optionally, the personal trajectory vector set is used for location prediction to predict the user's future location or movement path. Optionally, the personal trajectory vector set is used to analyze the user's behavior pattern to make personalized service recommendations or optimize network resource allocation. Optionally, the personal trajectory vector set is used for anomaly detection to detect abnormal patterns in user behavior, such as abnormal movement speed or position offset.

[0058] Exemplarily, the location data and signal strength of the user terminal at different time points are as follows:

[0059] time Base station 1 signal strength Base station 2 signal strength GPS Coordinates T1 -70dBm -65dBm (40.748,-73.985) T2 -68dBm -67dBm (40.749,-73.986) T3 -72dBm -66dBm (40.750,-73.987)

[0060] From these data, we can calculate the user's movement speed, frequently visited areas, signal strength changes, etc., and then generate the following personal trajectory vector:

[0061] Vector 1: contains the user's location and signal strength characteristics at time T1;

[0062] Vector 2: contains the user's position, movement speed, and signal strength change at time T2;

[0063] Vector 3: contains the user's location, movement direction, and dwell time at time T3.

[0064] Among them, each vector can be further used to analyze user behavior patterns, predict future locations, etc.

[0065] In a feasible embodiment, the spatial vector in the personal trajectory vector set is encoded to obtain time encoding and position encoding.

[0066] Time encoding can be generated by timestamp conversion, where the timestamp is converted into discrete time units, such as hours, minutes, and seconds. For example, the time of day is converted into "08:00" or "14:30". Alternatively, the time is encoded using sine and cosine functions, retaining the periodic characteristics of time. For example, sin(2π*t / 24) and cos(2π*t / 24) are used to encode hours. Time encoding includes time series features, where time intervals, such as the time it takes for a user to move from one location to another, are calculated as time series features.

[0067] Location encoding can be the standardization of location coordinates, that is, standardizing location coordinates (such as longitude and latitude) to a fixed range (0 to 1) to make them suitable for machine learning models. Location encoding can be dividing a geographic area into grids or regions, and each grid is given a unique code. For example, a city is divided into several grids, and the location is represented by the grid number. Location encoding can be the use of location embedding technology to map location coordinates to a low-dimensional vector space. This method can capture the similarities between geographic locations.

[0068] Combining time coding and location coding can create a full-time period personal trajectory vector containing time and location information, which can better capture the user's behavior patterns and location changes.

[0069] Step S20, pre-training the large base model based on the time code and the position code, and extracting the static features of the user and the static features of the space contained in the pre-trained large base model.

[0070] Optionally, a transformer model architecture based on a self-attention mechanism is selected as the base model, and the personal trajectory vector set is input for encoding to generate time encoding and position encoding.

[0071] Optionally, step S20 includes: constructing a pre-training task, the pre-training task including predicting missing grid identifiers and corresponding link times in the personal trajectory vector set, the missing grid identifiers being dynamically masked grid identifiers of a set proportion; based on the pre-training task, inputting the time code and the position code into the base large model for pre-training.

[0072] It should be noted that after generating the time code and position code, the bidirectional attention mechanism is used to input the time code and position code into the base large model, and the pre-training task of the dynamic large model is to predict the set proportion of dynamically masked grid IDs (Identification) and corresponding link times in each unlabeled trajectory, iterate and converge the model to form a pre-trained dynamic large model skeleton containing static and dynamic features of users and spaces. Optionally, the set proportion is 20%.

[0073] The occupational attributes of users determine that the dynamic link relationship between their mobile phones and base stations exhibits specific statistical laws or characteristics. A pre-trained dynamic learning large model based on the personal trajectory vector set is constructed. Based on time coding and position coding, long-sequence spatial observation learning with absolute time information is performed to obtain dynamic and static representations of users and regions.

[0074] Static features are relatively stable features of users over a period of time, and they are not likely to change over time. Static features include regular patterns in the link relationship between users and base stations, which can reflect the user's occupational attributes. For example, a person's occupation may affect their daily mobility patterns, such as the places they frequently visit, the time they visit these places, etc.

[0075] Exemplarily, static features include: Resident location, i.e., places frequently visited by users, such as home addresses or workplaces, which are relatively stable over a period of time. Visit frequency, such as the frequency of users visiting specific places. Visit time, the time pattern of users visiting specific places. Activity area, i.e., the main area of ​​user activity, such as some people may be mainly active in commercial areas, while others may be more active in residential areas. Mobility pattern, i.e., the pattern of users moving from one place to another, such as daily commuting routes. Base station connection pattern, i.e., the strength and duration of the user's connection with a specific base station, reflecting the user's stay time in a certain area. Device usage habits, i.e., the user's habits of using mobile phones or other mobile devices, such as the frequency of use of certain applications. Social network structure, i.e., the user's position in the social network, such as contacts, the frequency and intensity of contact with contacts. Consumption habits, such as the consumption pattern of users in specific places. The user's daily activity pattern, such as the time to get up, the time to go to bed, and the time for leisure activities.

[0076] Step S30, training a pre-built occupation classifier according to the static features of the user and the static features of the space to obtain a trained occupation classifier; the trained occupation classifier is used to identify the user's occupation.

[0077] Optionally, based on the long-term invariant characteristics of the user's occupational attributes, static features of the user and space contained in the hidden layer of the generated pre-trained dynamic large model are selected to construct an occupation classifier, wherein the occupation classifier can be a binary classifier or a multi-classifier.

[0078] Optionally, training the classifier in step S30 includes: obtaining a true value sample set of occupations, and constructing a training trajectory vector set based on the true value sample set of occupations; generating time coding and position coding of the training trajectory vector set; and training a pre-constructed occupation classifier based on the time coding and position coding of the training trajectory vector set, as well as the static features of the user and the static features of the space.

[0079] It should be noted that after building the classifier, based on the long-term true value sample data of the occupation, the true value of the occupation that has been engaged in the occupation for a longer period of time is selected as much as possible, and the personal trajectory vector of the entire time period is constructed, and a large model classifier mode with full parameter fine-tuning based on the long-term true value samples is trained.

[0080] Specifically, the truth value of a person who has been engaged in the occupation for a longer period of time is selected from the user's occupational truth value sample set, and the personal trajectory vector set is constructed based on the selected truth value. After time encoding and position encoding, the classifier is trained to trigger the full parameter fine-tuning of the pre-trained dynamic large model skeleton to refresh the hidden user's static features, space static features and classifier parameters, so as to better capture the corresponding spatiotemporal characteristics of people with different occupational types.

[0081] Optionally, the personal trajectory vector set corresponding to the user terminal is input into a trained classifier to obtain a recognition result of the classifier on the user occupation of the user terminal.

[0082] In the technical solution of this embodiment, according to the dynamic link relationship between the user terminal and the base station, a personal trajectory vector set corresponding to the user terminal is generated, and the time code and position code of the personal trajectory vector set are generated; the base large model is pre-trained based on the time code and position code, and the static features of the user and the static features of the space contained in the pre-trained base large model are extracted; according to the static features of the user and the static features of the space, the pre-constructed occupation classifier is trained to obtain a trained occupation classifier; the trained occupation classifier is used for user occupation identification. The personal trajectory vector set is generated through the dynamic link relationship to reduce the calculation load of the large model, improve the comprehensiveness of the personal trajectory determination, and facilitate the subsequent user occupation identification. The base large model is pre-trained according to the time code and position code to obtain static features to construct a classifier, and the user's occupation is identified by the classifier constructed by the static features shown in the dynamic link relationship between the user and the base station, thereby improving the accuracy of user occupation identification.

[0083] Based on the first embodiment of the present application, in the second embodiment of the present application, the same or similar contents as those in the above-mentioned embodiment 1 can be referred to the above introduction, and will not be repeated in the following. Figure 2 , step S10 comprises:

[0084] Step S11, generating a ternary grid group of the user terminal according to the dynamic link relationship between the user terminal and the base station, the ternary grid group including the terminal number, the grid identifier of the grid area where the base station is located, and the time when the user terminal is in the grid identifier;

[0085] Step S12: generating a personal trajectory vector set with a time stamp according to the triple grid group.

[0086] Optionally, before step S11, the method further includes: acquiring a target area; dividing the target area into grids to form a grid set covering the target area, wherein the grid set includes at least one grid area, and each grid area corresponds to a grid identifier.

[0087] For the target area, which includes the urban area, the target area is divided into grids of preset sizes, for example, the preset size is 300 meters * 300 meters, to form a grid set covering the target area, and each grid in the grid set has a unique grid identifier.

[0088] According to the dynamic link relationship between the terminal and the base station, the city is appropriately gridded and transformed into a new ternary grid group, forming a personal trajectory vector set data with a timestamp.

[0089] Optionally, step S11 also includes: generating a terminal number, a base station longitude and latitude, and a time when the user terminal is at the base station longitude and latitude according to the dynamic link relationship between the user terminal and the base station; determining a grid identifier where the base station is located according to the base station longitude and latitude; generating a ternary grid group of the user terminal according to the grid identifier where the base station is located, the terminal number, and the time. Exemplarily, based on the dynamic link relationship between the mobile phone and the base station, a ternary is formed, wherein the ternary includes the mobile phone number, the base station longitude and latitude, and the time. Among them, according to the base station longitude and latitude, the grid where the base station is located is calculated, and its grid identifier is obtained, which is converted to form a new ternary grid group, that is, the ternary grid group includes the mobile phone number, the grid identifier, and the time.

[0090] Select appropriate sampling period and sampling frequency, such as a 1-month period and a frequency of every 15 minutes, and construct a personal trajectory vector set based on a ternary grid group to form a data set suitable for pre-training large models and reduce the computational load of large models.

[0091] The massive data based on the dynamic link relationship between the user's mobile phone terminal and the base station contains the dynamic and static characteristics of the user and the dynamic and static characteristics of the space, including the specific statistical laws or characteristics of the link relationship between a person and the base station determined by his or her occupational attributes. Through the new ternary spatiotemporal grid group, the pre-trained dynamic large model is generated, and the prediction of the appropriate proportion of dynamically masked grid identifiers and corresponding time in each unlabeled trajectory is used as the pre-training task, and the model is iterated and converged. Finally, based on the pre-trained dynamic large model formed, the occupational truth value is used to drive the full amount of fine-tuning of the parameters of the pre-trained dynamic learning large model to form a multi-classification model, achieve the goal of one model serving multiple tasks, and alleviate the shortcomings of the traditional large model that considers less explicit time factors, thereby solving or alleviating the difficulties of traditional methods in data collection, complex feature vector construction, long time period, and high cost.

[0092] The transformer model architecture based on the self-attention mechanism is selected as the base model. In the target area or city, a grid set of appropriate size is set to form full coverage, in which each small grid has a unique identifier. Then, a triple is formed through the dynamic link relationship between the mobile phone and the base station, and the longitude and latitude of the base station are converted into the corresponding grid identifier to form a new triple grid group. According to the appropriate sampling time, a personal trajectory vector composed of the triple grid group is formed. After all personal trajectory vector sets are time-encoded and position-encoded, they are input into the base model using the bidirectional attention mechanism to generate a pre-trained dynamic large model. The pre-training task is to predict the appropriate proportion of dynamically masked grid identifiers and corresponding times in each unlabeled trajectory, iterate and converge the model, and finally, based on the static features of users and spaces contained in the hidden layer of the pre-trained dynamic large model, the long-period true value sample set is used to fine-tune the full parameters of the pre-trained dynamic large model to refresh the static features of users and spaces, so as to better capture the corresponding features of people of different occupational types, thereby constructing a binary or multi-classifier to achieve accurate prediction of individual occupations.

[0093] In the technical solution of this embodiment, a ternary grid group is constructed through a dynamic link relationship to generate a personal trajectory vector set, which is input into a large model to reduce the calculation load of the large model and improve the comprehensiveness of personal trajectory determination to facilitate subsequent user occupation identification.

[0094] It should be noted that the above examples are only used to understand the present application and do not constitute a limitation on the user occupation identification method of the present application. More simple transformations based on this technical concept are all within the scope of protection of the present application.

[0095] This application also provides a user occupation identification device, please refer to Figure 3 , the user occupation identification device comprises:

[0096] A generating module 10 is used to generate a personal trajectory vector set corresponding to the user terminal according to the dynamic link relationship between the user terminal and the base station, and to generate a time code and a position code of the personal trajectory vector set;

[0097] An extraction module 20, configured to pre-train the base large model based on the time code and the position code, and extract static features of the user and static features of the space contained in the pre-trained base large model;

[0098] The training module 30 is used to train a pre-built occupation classifier according to the static features of the user and the static features of the space to obtain a trained occupation classifier; the trained occupation classifier is used to identify the user's occupation.

[0099] Optionally, the step of generating a personal trajectory vector set corresponding to the user terminal according to the dynamic link relationship between the user terminal and the base station includes:

[0100] Generate a ternary grid group of the user terminal according to the dynamic link relationship between the user terminal and the base station, the ternary grid group including the terminal number, the grid identifier of the grid area where the base station is located, and the time when the user terminal is located at the grid identifier;

[0101] A set of personal trajectory vectors with timestamps is generated according to the triple grid group.

[0102] Optionally, before the step of generating a ternary mesh group of the user terminal according to the dynamic link relationship between the user terminal and the base station, the step further includes:

[0103] Get the target area;

[0104] The target area is divided into grids to form a grid set covering the target area, wherein the grid set includes at least one grid area, and each of the grid areas corresponds to a grid identifier.

[0105] Optionally, the step of generating a ternary mesh group of the user terminal according to the dynamic link relationship between the user terminal and the base station includes:

[0106] Generate the terminal number, the longitude and latitude of the base station and the time when the user terminal is at the longitude and latitude of the base station according to the dynamic link relationship between the user terminal and the base station;

[0107] Determine the grid identifier where the base station is located according to the latitude and longitude of the base station;

[0108] A ternary grid group of the user terminal is generated according to the grid identifier where the base station is located, the terminal number and the time.

[0109] Optionally, the step of training a pre-built occupation classifier according to the static features of the user and the static features of the space includes:

[0110] Obtaining a true value sample set of occupations, and constructing a training trajectory vector set according to the true value sample set of occupations;

[0111] Generating a time code and a position code of the training trajectory vector set;

[0112] A pre-built occupation classifier is trained based on the time coding and position coding of the training trajectory vector set, as well as the static features of the user and the static features of the space.

[0113] Optionally, the step of pre-training the base large model based on the time code and the position code includes:

[0114] Constructing a pre-training task, the pre-training task comprising predicting missing grid identifiers and corresponding link times in the personal trajectory vector set, the missing grid identifiers being dynamically masked grid identifiers of a set proportion;

[0115] Based on the pre-training task, the time code and the position code are input into the base large model for pre-training.

[0116] The user occupation identification device provided by the present application adopts the user occupation identification method in the above embodiment, which can solve the technical problem of low accuracy in predicting the user occupation. Compared with the prior art, the beneficial effects of the user occupation identification device provided by the present application are the same as the beneficial effects of the user occupation identification method provided by the above embodiment, and other technical features in the user occupation identification device are the same as the features disclosed in the above embodiment method, which will not be repeated here.

[0117] The present application provides a user occupation identification device, which includes: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the user occupation identification method in the above-mentioned embodiment one.

[0118] Reference below Figure 4 , which shows a schematic diagram of the structure of a user occupation identification device suitable for implementing the embodiment of the present application. The user occupation identification device in the embodiment of the present application may include but is not limited to mobile terminals such as mobile phones, laptop computers, digital broadcast receivers, PDAs (Personal Digital Assistants), PADs (Portable Application Descriptions), PMPs (Portable Media Players), vehicle-mounted terminals (such as vehicle-mounted navigation terminals), etc., and fixed terminals such as digital TVs (Televisions), desktop computers, etc. Figure 4 The user occupation identification device shown is merely an example and should not bring any limitation to the functions and scope of use of the embodiments of the present application.

[0119] like Figure 4As shown, the user occupation identification device may include a processing device 1001 (e.g., a central processing unit, a graphics processor, etc.), which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM: Read Only Memory) 1002 or a program loaded from a storage device 1003 to a random access memory (RAM: Random Access Memory) 1004. In RAM1004, various programs and data required for the operation of the user occupation identification device are also stored. The processing device 1001, ROM1002, and RAM1004 are connected to each other through a bus 1005. An input / output (I / O) interface 1006 is also connected to the bus. Generally, the following systems can be connected to the I / O interface 1006: an input device 1007 including, for example, a touch screen, a touch pad, a keyboard, a mouse, an image sensor, a microphone, an accelerometer, a gyroscope, etc.; an output device 1008 including, for example, a liquid crystal display (LCD: Liquid Crystal Display), a speaker, a vibrator, etc.; a storage device 1003 including, for example, a magnetic tape, a hard disk, etc.; and a communication device 1009. The communication device 1009 can allow the user occupation identification device to communicate with other devices wirelessly or by wire to exchange data. Although the figure shows a user occupation identification device with various systems, it should be understood that it is not required to implement or have all the systems shown. More or fewer systems can be implemented or provided alternatively.

[0120] In particular, according to the embodiments disclosed in the present application, the process described above with reference to the flowchart can be implemented as a computer software program. For example, the embodiments disclosed in the present application include a computer program product, which includes a computer program carried on a computer-readable medium, and the computer program includes a program code for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from a network through a communication device, or installed from a storage device 1003, or installed from a ROM 1002. When the computer program is executed by the processing device 1001, the above-mentioned functions defined in the method of the embodiment disclosed in the present application are executed.

[0121] The user occupation identification device provided by the present application adopts the user occupation identification method in the above embodiment, which can solve the technical problem of low accuracy in predicting the user occupation. Compared with the prior art, the beneficial effects of the user occupation identification device provided by the present application are the same as the beneficial effects of the user occupation identification method provided by the above embodiment, and the other technical features in the user occupation identification device are the same as the features disclosed in the method of the previous embodiment, which will not be repeated here.

[0122] It should be understood that the various parts disclosed in this application can be implemented by hardware, software, firmware or a combination thereof. In the description of the above embodiments, specific features, structures, materials or characteristics can be combined in any one or more embodiments or examples in a suitable manner.

[0123] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art who is familiar with the present technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

[0124] The present application provides a computer-readable storage medium having computer-readable program instructions (ie, computer programs) stored thereon, and the computer-readable program instructions are used to execute the user occupation identification method in the above-mentioned embodiment.

[0125] The computer-readable storage medium provided in the present application may be, for example, a USB flash drive, but is not limited to electrical, magnetic, optical, electromagnetic, infrared, or semiconductor systems, systems or devices, or any combination of the above. More specific examples of computer-readable storage media may include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In this embodiment, the computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in combination with an instruction execution system, system or device. The program code contained on the computer-readable storage medium may be transmitted using any appropriate medium, including but not limited to: wires, optical cables, RF (Radio Frequency), etc., or any suitable combination of the above.

[0126] The computer-readable storage medium may be included in the user occupation identification device; or may exist independently without being assembled into the user occupation identification device.

[0127] The above-mentioned computer-readable storage medium carries one or more programs. When the above-mentioned one or more programs are executed by the user occupation identification device, the user occupation identification device: generates a personal trajectory vector set through a dynamic link relationship to reduce the calculation load of the large model, improves the comprehensiveness of the personal trajectory determination, and facilitates the subsequent user occupation identification, pre-trains the base large model according to time coding and position coding to obtain static features to construct a classifier, and identifies the user's occupation through the classifier constructed by the static features shown in the dynamic link relationship between the user and the base station, thereby improving the accuracy of user occupation identification.

[0128] Computer program code for performing the operations of the present application may be written in one or more programming languages ​​or a combination thereof, including object-oriented programming languages ​​such as Java, Smalltalk, C++, and conventional procedural programming languages ​​such as "C" or similar programming languages. The program code may be executed entirely on the user's computer, partially on the user's computer, as a separate software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In the case of a remote computer, the remote computer may be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0129] The flow chart and block diagram in the accompanying drawings illustrate the possible architecture, function and operation of the system, method and computer program product according to various embodiments of the present application. In this regard, each square box in the flow chart or block diagram can represent a module, a program segment or a part of a code, and the module, the program segment or a part of the code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the square box can also occur in a sequence different from that marked in the accompanying drawings. For example, two square boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each square box in the block diagram and / or flow chart, and the combination of the square boxes in the block diagram and / or flow chart can be implemented with a dedicated hardware-based system that performs a specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions.

[0130] The modules involved in the embodiments described in this application may be implemented by software or hardware, wherein the name of the module does not constitute a limitation on the unit itself in some cases.

[0131] The readable storage medium provided in the present application is a computer-readable storage medium, which stores computer-readable program instructions (i.e., computer programs) for executing the above-mentioned user occupation identification method, and can solve the technical problem of low accuracy in predicting user occupations. Compared with the prior art, the beneficial effects of the computer-readable storage medium provided in the present application are the same as the beneficial effects of the user occupation identification method provided in the above-mentioned embodiment, and will not be elaborated here.

[0132] The present application also provides a computer program product, including a computer program, which implements the steps of the above-mentioned user occupation identification method when executed by a processor.

[0133] The computer program product provided by the present application can solve the technical problem of low accuracy in predicting user occupations. Compared with the prior art, the beneficial effects of the computer program product provided by the present application are the same as those of the user occupation identification method provided by the above embodiment, which will not be described in detail here.

[0134] The above descriptions are only some embodiments of the present application, and are not intended to limit the patent scope of the present application. All equivalent structural changes made using the contents of the present application specification and drawings under the technical concept of the present application, or direct / indirect applications in other related technical fields are included in the patent protection scope of the present application.

Claims

1. A method for identifying a user's occupation, characterized in that: The method includes: Generate a personal trajectory vector set corresponding to the user terminal according to the dynamic link relationship between the user terminal and the base station, and generate a time code and a position code of the personal trajectory vector set; Pre-training the base large model based on the time code and the position code, and extracting static features of the user and static features of the space contained in the pre-trained base large model; According to the static features of the user and the static features of the space, a pre-constructed occupation classifier is trained to obtain a trained occupation classifier; the trained occupation classifier is used to identify the user's occupation.

2. The method according to claim 1, characterized in that The step of generating a personal trajectory vector set corresponding to the user terminal according to the dynamic link relationship between the user terminal and the base station comprises: Generate a ternary grid group of the user terminal according to the dynamic link relationship between the user terminal and the base station, the ternary grid group including the terminal number, the grid identifier of the grid area where the base station is located, and the time when the user terminal is located at the grid identifier; A set of personal trajectory vectors with timestamps is generated according to the triple grid group.

3. The method according to claim 2, characterized in that Before the step of generating a ternary mesh group of the user terminal according to the dynamic link relationship between the user terminal and the base station, the method further includes: Get the target area; The target area is divided into grids to form a grid set covering the target area, wherein the grid set includes at least one grid area, and each of the grid areas corresponds to a grid identifier.

4. The method according to claim 3, characterized in that The step of generating a ternary mesh group of the user terminal according to the dynamic link relationship between the user terminal and the base station comprises: Generate the terminal number, the longitude and latitude of the base station and the time when the user terminal is at the longitude and latitude of the base station according to the dynamic link relationship between the user terminal and the base station; Determine the grid identifier where the base station is located according to the latitude and longitude of the base station; A ternary grid group of the user terminal is generated according to the grid identifier where the base station is located, the terminal number and the time.

5. The method according to claim 1, characterized in that The step of training the pre-built occupation classifier according to the static features of the user and the static features of the space comprises: Obtaining a true value sample set of occupations, and constructing a training trajectory vector set according to the true value sample set of occupations; Generating a time code and a position code of the training trajectory vector set; A pre-built occupation classifier is trained based on the time coding and position coding of the training trajectory vector set, as well as the static features of the user and the static features of the space.

6. The method according to claim 1, characterized in that The step of pre-training the base large model based on the time code and the position code comprises: Constructing a pre-training task, the pre-training task comprising predicting missing grid identifiers and corresponding link times in the personal trajectory vector set, the missing grid identifiers being dynamically masked grid identifiers of a set proportion; Based on the pre-training task, the time code and the position code are input into the base large model for pre-training.

7. A user occupation identification device, characterized in that: The device comprises: A generating module, configured to generate a personal trajectory vector set corresponding to the user terminal according to a dynamic link relationship between the user terminal and the base station, and to generate a time code and a position code of the personal trajectory vector set; An extraction module, used for pre-training the base large model based on the time code and the position code, and extracting the static features of the user and the static features of the space contained in the pre-trained base large model; The training module is used to train a pre-built occupation classifier according to the static features of the user and the static features of the space to obtain a trained occupation classifier; the trained occupation classifier is used to identify the user's occupation.

8. A user occupation identification device, characterized in that: The device comprises: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the computer program is configured to implement the steps of the user occupation identification method according to any one of claims 1 to 6.

9. A storage medium, characterized in that: The storage medium is a computer-readable storage medium, and a computer program is stored on the storage medium. When the computer program is executed by a processor, the steps of the user occupation identification method according to any one of claims 1 to 6 are implemented.

10. A computer program product, characterized in that The computer program product comprises a computer program, and when the computer program is executed by a processor, the steps of the method for identifying a user's occupation as claimed in any one of claims 1 to 6 are implemented.

Images