User portrait updating method, device, equipment and storage medium

Abstract

The present application discloses a user portrait updating method, device, equipment and storage medium, the method comprising: obtaining offline data collected within a corresponding time period at a first time node, and updating a first offline user portrait generated at a second time node according to the offline data; in the case of receiving real-time data, adding the real-time data to a first temporary table, and determining whether the offline data is completely processed; if the offline data is not completely processed, generating a second temporary table according to the unprocessed data in the offline data, and updating the second offline user portrait according to the second temporary table and the first temporary table to obtain a first real-time user portrait; if the offline data is completely processed, updating the third offline user portrait according to the first temporary table to obtain a second real-time user portrait. Through the above technical means, the problem that the user portrait in the prior art cannot accurately predict the real-time needs of users is solved, and the service quality of the enterprise is guaranteed.

Description

Technical Field

[0001] The present application relates to the field of big data technology, and in particular to a user portrait updating method, device, equipment and storage medium. Background Art

[0002] User portraits are an effective tool for outlining target users, linking user demands and design directions. In the fields of the Internet, advertising, e-commerce, and finance, user portraits are used for personalized recommendations, precision marketing, and risk control, helping companies better meet user needs and improve service quality and efficiency. Companies can collect user data and analyze it to build user portraits.

[0003] In the prior art, user portraits are constructed based on pre-collected offline user data, that is, the user's historical behavior data is first collected, and the historical behavior data is analyzed and processed to construct the user portrait. However, as the speed of Internet content updates and iterations continues to accelerate, companies are unable to accurately predict users' real-time needs based on offline user portraits, which affects the company's service quality. Summary of the invention

[0004] The present application provides a user portrait updating method, device, equipment and storage medium, so that when real-time data arrives, the offline user portrait obtained by offline update is updated in real time through real-time data to obtain a real-time user portrait, so as to accurately predict the real-time needs of the user based on the real-time user portrait, thereby solving the problem in the prior art that the user portrait cannot accurately predict the real-time needs of the user, and ensuring the service quality of the enterprise.

[0005] In a first aspect, the present application provides a method for updating a user portrait, comprising:

[0006] Acquire offline data collected in a corresponding time period at a first time node, and update a first offline user portrait generated at a second time node according to the offline data, wherein the time interval between the second time node and the first time node is the time period corresponding to the first time node;

[0007] When receiving the real-time data, adding the real-time data to the first temporary table, and determining whether the offline data is completely processed;

[0008] If the offline data is not completely processed, a second temporary table is generated according to the unprocessed data in the offline data, and a second offline user portrait is updated according to the second temporary table and the first temporary table to obtain a first real-time user portrait, and the second offline user portrait is obtained by updating the first offline user portrait based on the processed data in the offline data;

[0009] If the offline data is completely processed, the third offline user portrait is updated according to the first temporary table to obtain a second real-time user portrait, and the third offline user portrait is obtained by updating the first offline user portrait based on the offline data.

[0010] In a second aspect, the present application provides a user portrait updating device, comprising:

[0011] A first offline updating module is configured to obtain offline data collected in a corresponding time period at a first time node, and update a first offline user portrait generated at a second time node according to the offline data, wherein the time interval between the second time node and the first time node is the time period corresponding to the first time node;

[0012] an update detection module, configured to, upon receiving the real-time data, add the real-time data to the first temporary table and determine whether the offline data is completely processed;

[0013] A first real-time update module is configured to generate a second temporary table according to the unprocessed data in the offline data if the offline data is not completely processed, and update the second offline user portrait according to the second temporary table and the first temporary table to obtain a first real-time user portrait, wherein the second offline user portrait is obtained by updating the first offline user portrait based on the processed data in the offline data;

[0014] The second real-time update module is configured to update the third offline user portrait according to the first temporary table to obtain a second real-time user portrait if the offline data is completely processed, and the third offline user portrait is obtained by updating the first offline user portrait based on the offline data.

[0015] In a third aspect, the present application provides a user portrait updating device, including:

[0016] one or more processors;

[0017] A memory stores 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 user portrait updating method as described in the first aspect.

[0018] In a fourth aspect, the present application provides a storage medium comprising computer executable instructions, which, when executed by a computer processor, are used to execute the user portrait updating method as described in the first aspect.

[0019] In the present application, offline data collected in a corresponding time period is obtained at a first time node, and the first offline user portrait generated at a second time node is updated according to the offline data; when real-time data is received, the real-time data is added to the first temporary table, and it is determined whether the offline data is completely processed; if the offline data is not completely processed, a second temporary table is generated according to the unprocessed data in the offline data, and the second offline user portrait is updated according to the second temporary table and the first temporary table to obtain the first real-time user portrait; if the offline data is completely processed, the third offline user portrait is updated according to the first temporary table to obtain the second real-time user portrait. Through the above technical means, the user portrait is updated offline at a fixed time node according to the offline data collected in a fixed time period, and when the real-time data arrives, the user portrait obtained by the offline update is updated in real time by the real-time data to obtain the real-time user screen, so that the real-time user portrait can reflect the current user characteristics and maintain consistency with the historical user characteristics, and the real-time needs of the user can be accurately predicted according to the real-time user portrait, which solves the problem that the user portrait in the prior art cannot accurately predict the real-time needs of the user, and ensures the service quality of the enterprise. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] Figure 1 This is a flow chart of a user portrait updating method provided in an embodiment of the present application;

[0021] Figure 2 is a flow chart of offline updating of a first offline user portrait provided by an embodiment of the present application;

[0022] Figure 3 is a flow chart of updating the second offline user portrait in real time provided by an embodiment of the present application;

[0023] Figure 4 is a flow chart of offline updating of a second offline user portrait provided by an embodiment of the present application;

[0024] Figure 5 This is a flow chart of updating the third offline user portrait in real time provided by an embodiment of the present application;

[0025] Figure 6 It is a structural schematic diagram of a user portrait updating device provided in an embodiment of the present application;

[0026] Figure 7 It is a structural diagram of a user portrait updating device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0027] In order to make the purpose, technical scheme and advantages of the present application clearer, the specific embodiments of the present application are further described in detail below in conjunction with the accompanying drawings. It is understood that the specific embodiments described herein are only used to explain the present application, rather than to limit the present application. It should also be noted that, for the convenience of description, only the part related to the present application but not all the contents are shown in the accompanying drawings. Before discussing the exemplary embodiments in more detail, it should be mentioned that some exemplary embodiments are described as processes or methods depicted as flow charts. Although the flow chart describes each operation (or step) as a sequential process, many of the operations therein can be implemented in parallel, concurrently or simultaneously. In addition, the order of each operation can be rearranged. The process can be terminated when its operation is completed, but it can also have additional steps not included in the accompanying drawings. The process can correspond to a method, a function, a procedure, a subroutine, a subprogram, etc.

[0028] The terms "first", "second", etc. in the specification and claims of the present application are used to distinguish similar objects, and are not used to describe a specific order or sequence. It should be understood that the data used in this way can be interchangeable under appropriate circumstances, so that the embodiments of the present application can be implemented in an order other than those illustrated or described here, and the objects distinguished by "first", "second", etc. are generally of one type, and the number of objects is not limited. For example, the first object can be one or more. In addition, "and / or" in the specification and claims represents at least one of the connected objects, and the character " / " generally indicates that the objects associated with each other are in an "or" relationship.

[0029] In the more common existing implementation method, user portraits are built based on pre-collected offline user data, that is, the user's historical behavior data is first collected, and the historical behavior data is analyzed and processed to build user portraits. However, as the update and iteration speed of Internet content continues to accelerate, companies cannot accurately predict users' real-time needs based on offline user portraits, which affects the company's service quality.

[0030] To solve the above problem, this embodiment provides a user portrait updating method, so that when real-time data arrives, the offline user portrait obtained by offline update is updated in real time through real-time data to obtain a real-time user portrait, so as to accurately predict the real-time needs of the user according to the real-time user portrait and ensure the service quality of the enterprise.

[0031] The user portrait updating method provided in this embodiment can be executed by a user portrait updating device, which can be implemented by software and / or hardware, and can be composed of two or more physical entities, or can be composed of one physical entity. For example, the user portrait updating device can be a server that provides various business services to users, such as an e-commerce server, a video server, a forum server, etc. These servers can predict user needs based on user portraits to recommend corresponding products, videos or posts to users based on user needs.

[0032] The user portrait updating device is installed with at least one type of operating system, wherein the operating system includes but is not limited to Android system, Linux system and Windows system. The user portrait updating device can install at least one application based on the operating system, and the application can be an application that comes with the operating system or an application downloaded from a third-party device or server. In this embodiment, the user portrait updating device has at least an application that can execute the user portrait updating method.

[0033] For ease of understanding, this embodiment is described by taking the e-commerce server as an example of the main body that executes the user portrait updating method.

[0034] Figure 1 A flowchart of a user portrait updating method provided by an embodiment of the present application is given. Figure 1 , the user portrait updating method specifically includes:

[0035] S110. Obtain offline data collected within a corresponding time period at a first time node, and update a first offline user portrait generated at a second time node according to the offline data. The time interval between the second time node and the first time node is the time period corresponding to the first time node.

[0036] In this embodiment, the e-commerce server can periodically update the user portrait offline, and the first time node and the second time node are respectively the two time nodes before and after the periodic offline update of the user portrait. For example, if the e-commerce server updates the user portrait offline at midnight every day, the e-commerce server can obtain the offline data collected on the previous day at midnight on the current day, and update the user portrait generated at midnight on the previous day based on the offline data. The current day at midnight is the first time node, the previous day at midnight is the second time node, and the one-day time period is the time period. The first offline user portrait is the user portrait obtained when the user portrait is updated offline at midnight on the previous day. This embodiment is described by taking the first time node as midnight on the current day and the second time node as midnight on the previous day as an example.

[0037] The offline data collected within the time period is the user data collected from the e-commerce client of the corresponding user within the corresponding time period. It can be understood that the offline data collected within the time period can reflect the basic information, behavioral habits and favorite preferences of the user within the time period, and the user portrait generated at the time node of the previous period can be updated offline based on the offline data collected within the time period. In this embodiment, in order to collect offline data, after the e-commerce client pushes the collected user data to the kafka cluster, the e-commerce server pulls the user data from the kafka cluster and saves the user data as offline data in the MYSQL database. When the current time reaches the first time node, the e-commerce server reads the offline data collected the previous day from the MYSQL database, and updates the first offline user portrait generated by the offline update at zero o'clock the previous day according to the offline data collected the previous day.

[0038] In one embodiment, the offline data includes the user's first behavior data and first content data. Exemplarily, the e-commerce client can generate corresponding first behavior data according to the browsing, clicking, searching, purchasing, analyzing, collecting and commenting behaviors input by the user. The e-commerce client can generate corresponding first content data according to the content browsed, clicked, searched, purchased, analyzed, collected and commented by the user, and push the first behavior data and the corresponding first content data to the kafka cluster in association. The e-commerce server pulls the associated cached first behavior data and the first content data from the kafka cluster, and saves the first behavior data and the first content data in association in the MYSQL database. Correspondingly, at midnight of the day, the e-commerce server obtains the first behavior data and the corresponding first content data collected the previous day from the MYSQL database, determines the user's behavior habits according to the first behavior data, and determines the user's product preferences according to the first content data and the first behavior data, so as to update the label of the first offline user portrait generated the previous day according to the behavior habits and product preferences.

[0039] In another embodiment, the offline data includes the user's first behavior data, first content data, first access path data, first access time data, first geographic location information and first device information. The user profile can be updated offline based on these data to analyze the user's characteristics from multiple levels and improve the accuracy of the user profile. Among them, the access path data is the access path when the user browses various contents, the access time data is the access time when the user browses various contents, the geographic location information is the geographic location of the user, and the first device information includes information such as the device type and device performance used by the user. Exemplary, Figure 2 1 is a flowchart of performing offline updating of the first offline user portrait provided by an embodiment of the present application. Figure 2 As shown, the step of updating the first offline user portrait specifically includes S1101-S1103:

[0040] S1101. Perform feature extraction on offline data to obtain first feature information, where the first feature information includes a first behavior feature, a first content feature, a first path feature, a first time feature, a first location feature, a first device feature, and a first statistical feature.

[0041] Exemplarily, feature extraction may be performed on the first behavior data to obtain a corresponding first behavior feature, feature extraction may be performed on the first content data to obtain a corresponding first content feature, feature extraction may be performed on the first access path data to obtain a corresponding first path feature, feature extraction may be performed on the first access time data to obtain a corresponding first time feature, feature extraction may be performed on the first geographic location information to obtain a corresponding first location feature, feature extraction may be performed on the first device information to obtain a corresponding first device feature, and statistics may be performed on the same content, the same behavior or the same path to obtain a corresponding first statistical feature.

[0042] S1102: Input the first feature information into a pre-trained first clustering model to obtain a first group category output by the first clustering model.

[0043] The first clustering model can be trained based on the pre-collected first feature information of various users and the corresponding group category labels, and the trained first clustering model has the function of predicting the group category to which the user belongs based on the first feature information. Exemplarily, the first behavior feature, the first content feature, the first path feature, the first time feature, the first location feature, the first device feature and the first statistical feature are input into the pre-trained first clustering model, and the first group category to which the user belongs is predicted by the first clustering model.

[0044] S1103: Determine a first user tag according to the first feature information and the first group category, and update the user tag of the first offline user portrait according to the first user tag.

[0045] Exemplarily, the first group category to which the user belongs is used as the first user tag, and the first user tag is added to the user tag of the first offline user portrait, so that the e-commerce server can subsequently recommend suitable products or purchase patterns to the user according to the first group category to which the user belongs. Further, the user's behavior habits, access path habits, and access time habits are determined according to the first behavior feature, the first path feature, the first time feature, and the first statistical feature, respectively, and the product preference is determined according to the first content feature, the first location feature, and the first device feature, and the habit and preference are added as the first user tag to the user tag of the first offline user portrait.

[0046] This embodiment performs feature analysis on the user's first behavior data, first content data, first access path data, first access time data, first geographic location information, and first device information to accurately determine user tags, so as to better understand user needs, optimize product strategies, and improve user experience.

[0047] S120: When real-time data is received, add the real-time data to the first temporary table, and determine whether the offline data is completely processed.

[0048] Exemplarily, after the e-commerce server starts to update the first offline user portrait offline at midnight of the same day, the user data collected by the e-commerce client after midnight of the same day is real-time data. For example, when the user uses the e-commerce client at 7 o'clock in the morning of the same day, the e-commerce client can collect the corresponding user data and push the user data to the kafka cluster. The e-commerce server pulls the user data from the kafka cluster and adds the user data as real-time data to the first temporary table. In combination with the above content, when the e-commerce server pulls the user data from the kafka cluster, the user data is saved in the MYSQL database as offline data for offline updating of the user portrait at midnight of the next day, and the user data is saved in the first temporary table as real-time data for real-time updating of the user portrait on that day.

[0049] It should be noted that the first temporary table and the MYSQL database have the same function, both of which are used to save the user data pulled by the e-commerce server from the Kafka cluster. In order to reduce the storage pressure of the MYSQL database and improve the efficiency of the e-commerce server in processing user data, the user data can be saved only in the first temporary table. At midnight the next day, the user data in the first temporary table is the user data collected the previous day. Therefore, the user data in the first temporary representation is used as offline data for offline updating of user portraits.

[0050] In addition, when the real-time data includes the user's second behavior data, second content data, second access path data, second access time data, second geographic location information, and second device information, the amount of real-time data is large. In order to improve the reading efficiency of the first temporary table, the first temporary table is generally cached in the memory, and the memory space is limited. When all the real-time data is cached in the memory, there may be a problem of insufficient memory space and affecting the system operation performance. In this regard, the second behavior data and the second content data can be obtained in the real-time data; the second behavior data and the second content data are associated and added to the first temporary table. It can be understood that the data stored in the first temporary table will be used to update the label of the user portrait in real time. Compared with other types of data, the second behavior data and the second content data can best represent the user's behavior habits and preferences. Therefore, the second behavior data and the second content data in the real-time data can be associated and saved in the first temporary table. Since only the second behavior data and the second content data are saved in the first temporary table, the storage space of the memory is greatly saved, the stability and efficiency of the system operation are guaranteed, and only the second behavior data and the second content data are processed later to update the user portrait in real time, without significantly reducing the accuracy of the user portrait, the real-time update efficiency of the user portrait is effectively improved.

[0051] Although offline updating of user portraits uses batch processing to process a large amount of offline data and update the labels of user portraits, when the amount of offline data is large or the real-time data arrives earlier, the offline data may not be fully processed, that is, some data in the offline data is not used to update the user portrait, that is, the updated user portrait at this time cannot accurately represent the user's habits and preferences. In this regard, this embodiment proposes that when the offline data has not been fully processed, the user portrait updated offline is updated in real time by combining the unprocessed data in the offline data and the real-time data, and when the offline data has been fully processed, the user portrait updated offline is updated in real time based on the real-time data, so as to ensure that the user portrait can reflect the current user characteristics and maintain consistency with the historical user characteristics, thereby ensuring the accuracy of the user portrait.

[0052] S130. If the offline data is not completely processed, a second temporary table is generated according to the unprocessed data in the offline data, and the second offline user portrait is updated according to the second temporary table and the first temporary table to obtain the first real-time user portrait. The second offline user portrait is obtained by updating the first offline user portrait based on the processed data in the offline data.

[0053] The second offline user portrait is the latest user portrait obtained by offline updating when the real-time data arrives. The first real-time user portrait is the user portrait obtained by updating the second offline user portrait in real time.

[0054] Exemplarily, the processed data and the unprocessed data in the offline data are distinguished by different identifiers, and the unprocessed data can be obtained according to the identifiers of each data in the offline data. The unprocessed data is added to the second temporary table to update the second offline user portrait in real time based on the user data stored in the second temporary table and the first temporary table.

[0055] In the case where the unprocessed data includes the user's first behavior data, first content data, first access path data, first access time data, first geographic location information, and first device information, the first behavior data and the corresponding first content data can be obtained from the unprocessed data, and the first behavior data and the corresponding first content data can be associated and added to the second temporary table. It can be understood that since the first temporary table and the second temporary table only store the user's behavior data and content data, while ensuring the consistency of data types, it can save memory storage space and improve the real-time update efficiency of user portraits.

[0056] In one embodiment, the user's behavioral habits can be determined based on the second behavior data in the first temporary table and the first behavior data in the second temporary table, and the user's product preferences can be determined based on the second behavior data and the corresponding second content data in the first temporary table and the first behavior data and the corresponding first content data in the second temporary table, and the behavioral habits and product preferences are used as new labels for the second offline user portrait.

[0057] In another embodiment, in order to more accurately analyze the user's label, the user's group category can be divided according to the user's behavior data and the corresponding content data, and the new label of the second offline user portrait can be updated according to the user's group category, behavior habits and product preferences. Figure 3 1 is a flowchart of updating the second offline user portrait in real time provided by an embodiment of the present application. Figure 3 As shown, the step of updating the second offline user portrait specifically includes S1301-S1303:

[0058] S1301. Extract a second behavior feature, a second content feature, and a second statistical feature according to the second behavior data and the second content data in the first temporary table and the first behavior data and the first content data in the second temporary table.

[0059] Exemplarily, feature extraction may be performed on the second behavior data in the first temporary table and the first behavior data in the second temporary table to obtain corresponding second behavior features, feature extraction may be performed on the second content data in the first temporary table and the first content data in the second temporary table to obtain corresponding second content features, statistics may be performed on the same behavior in the first behavior data and the second behavior data, and statistics may be performed on the same content in the first content data and the second content data to obtain corresponding second statistical features.

[0060] S1302: Input the second behavior feature, the second content feature, and the second statistical feature into a pre-trained second clustering model to obtain a second group category output by the second clustering model.

[0061] The second clustering model can be trained based on various pre-collected user behavior characteristics, content characteristics and statistical characteristics and corresponding group category labels, and the trained second clustering model has the function of predicting the group category to which the user belongs based on the behavior characteristics, content characteristics and statistical characteristics. Exemplarily, the second behavior characteristics, the second content characteristics and the second statistical characteristics are input into the pre-trained second clustering model, and the second group category to which the user belongs is predicted by the second clustering model.

[0062] S1303: Determine a second user tag according to the second group category, the second behavior feature, the second content feature, and the second statistical feature, and update the user tag of the second offline user portrait according to the second user tag to obtain a first real-time user portrait.

[0063] Exemplarily, the second group category to which the user belongs is used as the second user tag, and the second user tag is added to the user tag of the second offline user portrait, so that the e-commerce server can subsequently recommend suitable goods or purchase modes to the user according to the second group category to which the user belongs. Further, the user's behavior habits are determined according to the second behavior characteristics and the second statistical characteristics, and the user's product preferences are determined according to the second content characteristics and the second statistical characteristics, and the behavior habits and product preferences are added as the second user tag to the user tag of the second offline user portrait, so that the e-commerce server can subsequently recommend suitable browsing modes or purchase modes to the user in real time according to the user's behavior habits, and recommend suitable products to the user in real time according to the user's product preferences.

[0064] This embodiment improves the updating efficiency of user portraits while ensuring the accuracy of user portraits by performing feature and group analysis on offline unprocessed behavior data and content data as well as behavior data and content data acquired in real time, thereby ensuring the timeliness of real-time data, so that enterprises can recommend models and products that match users' current habits and preferences in real time based on updated user portraits, thereby improving users' usage experience.

[0065] It should be noted that although the first real-time user portrait is updated based on the unprocessed offline data and the real-time data received on the day to ensure the consistency of real-time data and historical data, the first real-time user portrait only analyzes the behavior data and user data when updating, so the accuracy is still lower than the analysis of multi-level user data. In this regard, this embodiment proposes that after the second offline user portrait is updated based on the second temporary table and the first temporary table to obtain the first real-time user portrait, the second offline user portrait can be updated offline using the unprocessed offline data to obtain a third offline user portrait. Exemplary, Figure 4 1 is a flowchart of performing offline updating of the second offline user portrait provided by an embodiment of the present application. Figure 4 As shown, the step of offline updating the second offline user portrait specifically includes S1501-S1502:

[0066] S1501. Perform feature extraction on unprocessed data to obtain second feature information, input the second feature information into a pre-trained first clustering model, and obtain a fourth group category output by the first clustering model.

[0067] Exemplarily, the second feature information includes a fourth behavior feature, a fourth content feature, a second path feature, a second time feature, a second location feature, a second device feature, and a fourth statistical feature. Feature extraction is performed on the first behavior data in the unprocessed data to obtain the fourth behavior feature, feature extraction is performed on the first content data in the unprocessed data to obtain the corresponding second content feature, feature extraction is performed on the first access path data in the unprocessed data to obtain the corresponding second path feature, feature extraction is performed on the first access time data in the unprocessed data to obtain the corresponding second time feature, feature extraction is performed on the first geographic location information in the unprocessed data to obtain the corresponding second location feature, feature extraction is performed on the first device information in the unprocessed data to obtain the corresponding second device feature, and statistics are performed on the same content, the same behavior, or the same path in the unprocessed data to obtain the corresponding fourth statistical feature.

[0068] The fourth behavior feature, the fourth content feature, the second path feature, the second time feature, the second location feature, the second device feature, and the fourth statistical feature are input into a pre-trained first clustering model, and the fourth group category to which the user belongs is predicted through the first clustering model.

[0069] S1502: Determine a fourth user tag according to the second feature information and the fourth group category, and update the user tag of the second offline user portrait according to the fourth user tag.

[0070] Exemplarily, the fourth group category to which the user belongs is used as the fourth user tag, and the fourth user tag is added to the user tag of the second offline user portrait, so that the e-commerce server can subsequently recommend suitable goods or purchase patterns to the user according to the fourth group category to which the user belongs. Further, the user's behavior habits, access path habits and access time habits are determined according to the fourth behavior feature, the second path feature, the second time feature and the fourth statistical feature, respectively, and the product preference is determined according to the fourth content feature, the second location feature and the second device feature, and the habit and preference are added as the fourth user tag to the user tag of the second offline user portrait. It should be noted that if the unprocessed data has been used to update the second offline user portrait when the next real-time data arrives, the third offline user portrait can be obtained, and then the third offline user portrait can be updated in real time based on the real-time data cached in the first temporary table. If the unprocessed data is still not used to update the second offline portrait when a real-time data arrives, then the latest updated second offline user portrait is updated in real time based on the remaining unprocessed offline data and the real-time data cached in the first temporary table.

[0071] This embodiment continues to update the second offline user portrait offline using unprocessed data during the idle interval between the end of the current real-time update and the arrival of the next real-time data, so as to obtain a more accurate offline user portrait, thereby improving the accuracy of the subsequently generated real-time user portrait.

[0072] S140: If the offline data is completely processed, the third offline user portrait is updated according to the first temporary table to obtain a second real-time user portrait, and the third offline user portrait is obtained by updating the first offline user portrait based on the offline data.

[0073] Among them, the third offline user portrait is a user portrait obtained by updating the first offline user portrait offline based on all offline data collected in the corresponding time period. It can be understood that before the real-time data arrives, the e-commerce server completes the task of updating the user portrait offline in advance to obtain the third offline user portrait. The second real-time user portrait is a user portrait obtained by updating the third offline user portrait in real time.

[0074] In one embodiment, the user's behavioral habits can be determined based on the second behavior data in the first temporary table, and the user's product preferences can be determined based on the second behavior data in the first temporary table and the corresponding second content data, and the behavioral habits and product preferences are used as new labels for the third offline user portrait.

[0075] In another embodiment, in order to more accurately analyze the user's label, the user's group category can be divided according to the user's behavior data and the corresponding content data, and the new label of the third offline user portrait can be updated according to the user's group category, behavior habits and product preferences. Figure 5 1 is a flowchart of updating the third offline user portrait in real time provided by an embodiment of the present application. Figure 5 As shown, the step of updating the third offline user portrait specifically includes S1401-S1403:

[0076] S1401. Extract a third behavior feature, a third content feature, and a third statistical feature according to the second behavior data and the second content data in the first temporary table.

[0077] Exemplarily, feature extraction may be performed on the second behavior data in the first temporary table to obtain corresponding third behavior features, feature extraction may be performed on the second content data in the first temporary table to obtain corresponding third content features, statistics may be performed on the same behavior in the first temporary table, and statistics may be performed on the same content in the first temporary table to obtain corresponding third statistical features.

[0078] S1402: Input the third behavior feature, the third content feature, and the third statistical feature into a pre-trained second clustering model to obtain a third group category output by the second clustering model.

[0079] Exemplarily, the third behavior feature, the third content feature, and the third statistical feature are input into a pre-trained second clustering model, and the second group category to which the user belongs is predicted by the second clustering model.

[0080] S1403: Determine a third user tag according to the third group category, the third behavior feature, the third content feature, and the third statistical feature, and update the user tag of the third offline user portrait according to the third user tag to obtain a second real-time user portrait.

[0081] Exemplarily, the third group category to which the user belongs is used as the third user tag, and the third user tag is added to the user tag of the third offline user portrait, so that the e-commerce server can subsequently recommend suitable goods or purchase modes to the user according to the third group category to which the user belongs. Further, the user's behavior habits are determined according to the third behavior characteristics and the third statistical characteristics, and the user's product preferences are determined according to the third content characteristics and the third statistical characteristics, and the behavior habits and product preferences are added as the third user tag to the user tag of the third offline user portrait, so that the e-commerce server can subsequently recommend suitable browsing modes or purchase modes to the user in real time according to the user's behavior habits, and recommend suitable products to the user in real time according to the user's product preferences.

[0082] This embodiment improves the updating efficiency of user portraits while ensuring the accuracy of user portraits by performing feature and group analysis on behavioral data and content data acquired in real time, thereby ensuring the timeliness of real-time data, so that enterprises can recommend models and products that match users' current habits and preferences in real time based on updated user portraits, thereby improving users' usage experience.

[0083] In summary, the user portrait updating method provided by the embodiment of the present application acquires offline data collected within a corresponding time period at a first time node, and updates the first offline user portrait generated at a second time node according to the offline data; when receiving real-time data, the real-time data is added to the first temporary table, and it is determined whether the offline data is completely processed; if the offline data is not completely processed, a second temporary table is generated according to the unprocessed data in the offline data, and the second offline user portrait is updated according to the second temporary table and the first temporary table to obtain the first real-time user portrait; if the offline data is completely processed, the third offline user portrait is updated according to the first temporary table to obtain the second real-time user portrait. Through the above technical means, the user portrait is updated offline at a fixed time node according to the offline data collected within a fixed time period, and when the real-time data arrives, the user portrait obtained by the offline update is updated in real time by the real-time data to obtain the real-time user screen, so that the real-time user portrait can reflect the current user characteristics and maintain consistency with the historical user characteristics, and the real-time needs of the user can be accurately predicted according to the real-time user portrait, which solves the problem that the user portrait in the prior art cannot accurately predict the real-time needs of the user, and ensures the service quality of the enterprise.

[0084] Based on the above embodiments, Figure 6 A schematic diagram of the structure of a user portrait updating device provided in an embodiment of the present application. Figure 6 The user portrait updating device provided in this embodiment specifically includes: a first offline updating module 21, an updating detection module 22, a first real-time updating module 23 and a second real-time updating module 24.

[0085] The first offline updating module 21 is configured to obtain offline data collected in a corresponding time period at a first time node, and update a first offline user portrait generated at a second time node according to the offline data, wherein the time interval between the second time node and the first time node is the time period corresponding to the first time node;

[0086] The update detection module 22 is configured to, upon receiving the real-time data, add the real-time data to the first temporary table and determine whether the offline data is completely processed;

[0087] The first real-time updating module 23 is configured to generate a second temporary table according to the unprocessed data in the offline data if the offline data is not completely processed, and update the second offline user portrait according to the second temporary table and the first temporary table to obtain the first real-time user portrait, and the second offline user portrait is obtained by updating the first offline user portrait based on the processed data in the offline data;

[0088] The second real-time updating module 24 is configured to update the third offline user portrait according to the first temporary table to obtain a second real-time user portrait if the offline data is completely processed. The third offline user portrait is obtained by updating the first offline user portrait based on the offline data.

[0089] On the basis of the above embodiment, the offline data includes the user's first behavior data, first content data, first access path data, first access time data, first geographic location information and first device information; accordingly, the first offline update module 21 includes: a first feature extraction unit, configured to perform feature extraction on the offline data to obtain first feature information, the first feature information includes a first behavior feature, a first content feature, a first path feature, a first time feature, a first location feature, a first device feature, and a first statistical feature; a first group classification unit, configured to input the first feature information into a pre-trained first clustering model to obtain a first group category output by the first clustering model; a first offline update unit, configured to determine a first user label according to the first feature information and the first group category, and update the user label of the first offline user portrait according to the first user label.

[0090] Based on the above embodiment, the real-time data includes the user's second behavior data, second content data, second access path data, second access time data, second geographic location information and second device information; accordingly, the update detection module 22 includes: a data acquisition unit, configured to acquire the second behavior data and the second content data in the real-time data; a first temporary table generation unit, configured to associate the second behavior data and the second content data and add them to the first temporary table.

[0091] Based on the above embodiment, the first real-time update module 23 includes: a second temporary table generating unit, configured to obtain first behavior data and corresponding first content data in unprocessed data, and associate the first behavior data and the corresponding first content data to add to the second temporary table.

[0092] On the basis of the above embodiment, the first real-time update module 23 includes: a second feature extraction unit, configured to extract the second behavior feature, the second content feature and the second statistical feature according to the second behavior data and the second content data in the first temporary table and the first behavior data and the first content data in the second temporary table; a second group classification unit, configured to input the second behavior feature, the second content feature and the second statistical feature into a pre-trained second clustering model to obtain a second group category output by the second clustering model; a first real-time update unit, configured to determine a second user label according to the second group category, the second behavior feature, the second content feature and the second statistical feature, and update the user label of the second offline user portrait according to the second user label to obtain a first real-time user portrait.

[0093] On the basis of the above embodiment, the second real-time update module 24 includes: a third feature extraction unit, configured to extract a third behavior feature, a third content feature and a third statistical feature according to the second behavior data and the second content data in the first temporary table; a third group classification unit, configured to input the third behavior feature, the third content feature and the third statistical feature into a pre-trained second clustering model to obtain a third group category output by the second clustering model; a second real-time update unit, configured to determine a third user label according to the third group category, the third behavior feature, the third content feature and the third statistical feature, and update the user label of the third offline user portrait according to the third user label to obtain a second real-time user portrait.

[0094] On the basis of the above embodiment, the user portrait updating device also includes a second offline updating module, which includes: a fourth group classification unit, which is configured to perform feature extraction on unprocessed data to obtain second feature information after updating the second offline user portrait according to the second temporary table and the first temporary table to obtain the first real-time user portrait, and input the second feature information into the pre-trained first clustering model to obtain the fourth group category output by the first clustering model; the second offline updating unit is configured to determine a fourth user tag according to the second feature information and the fourth group category, and update the user tag of the second offline user portrait according to the fourth user tag.

[0095] As mentioned above, the user portrait updating device provided by the embodiment of the present application acquires offline data collected in a corresponding time period at a first time node, and updates the first offline user portrait generated at a second time node according to the offline data; when receiving real-time data, the real-time data is added to the first temporary table, and it is determined whether the offline data is completely processed; if the offline data is not completely processed, a second temporary table is generated according to the unprocessed data in the offline data, and the second offline user portrait is updated according to the second temporary table and the first temporary table to obtain the first real-time user portrait; if the offline data is completely processed, the third offline user portrait is updated according to the first temporary table to obtain the second real-time user portrait. Through the above technical means, the user portrait is updated offline according to the offline data collected in a fixed time period at a fixed time node, and when the real-time data arrives, the user portrait obtained by the offline update is updated in real time by the real-time data to obtain the real-time user screen, so that the real-time user portrait can reflect the current user characteristics and keep consistency with the historical user characteristics, and the real-time needs of the user can be accurately predicted according to the real-time user portrait, which solves the problem that the user portrait in the prior art cannot accurately predict the real-time needs of the user, and ensures the service quality of the enterprise.

[0096] The user portrait updating device provided in the embodiment of the present application can be used to execute the user portrait updating method provided in the above embodiment, and has corresponding functions and beneficial effects.

[0097] Figure 7 This is a schematic diagram of the structure of a user portrait updating device provided in an embodiment of the present application, with reference to Figure 7 The user portrait updating device includes: a processor 31, a memory 32, a communication device 33, an input device 34, and an output device 35. The number of processors 31 in the user portrait updating device can be one or more, and the number of memories 32 in the user portrait updating device can be one or more. The processor 31, memory 32, communication device 33, input device 34, and output device 35 of the user portrait updating device can be connected via a bus or other means.

[0098] The memory 32, as a computer-readable storage medium, can be used to store software programs, computer executable programs and modules, such as program instructions / modules corresponding to the user portrait update method of any embodiment of the present application (for example, the first offline update module 21, the update detection module 22, the first real-time update module 23 and the second real-time update module 24 in the user portrait update device). The memory 32 may mainly include a program storage area and a data storage area, wherein the program storage area may store an operating system, an application required for at least one function; the data storage area may store data created according to the use of the device, etc. In addition, the memory 32 may include a high-speed random access memory, and may also include a non-volatile memory, such as at least one disk storage device, a flash memory device, or other non-volatile solid-state storage device. In some instances, the memory may further include a memory remotely arranged relative to the processor, and these remote memories may be connected to the device via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network and a combination thereof.

[0099] The communication device 33 is used for data transmission.

[0100] The processor 31 executes various functional applications and data processing of the device by running software programs, instructions and modules stored in the memory 32, that is, implements the above-mentioned user portrait updating method.

[0101] The input device 34 may be used to receive input digital or character information and generate key signal input related to user settings and function control of the device. The output device 35 may include a display device such as a display screen.

[0102] The user portrait updating device provided above can be used to execute the user portrait updating method provided in the above embodiment, and has corresponding functions and beneficial effects.

[0103] An embodiment of the present application also provides a storage medium containing computer executable instructions, which are used to execute a user portrait updating method when executed by a computer processor. The user portrait updating method includes: acquiring offline data collected within a corresponding time period at a first time node, updating a first offline user portrait generated at a second time node according to the offline data, and the time interval between the second time node and the first time node is the time period corresponding to the first time node; when real-time data is received, adding the real-time data to a first temporary table, and determining whether the offline data is completely processed; if the offline data is not completely processed, generating a second temporary table according to the unprocessed data in the offline data, updating the second offline user portrait according to the second temporary table and the first temporary table to obtain a first real-time user portrait, and the second offline user portrait is obtained by updating the first offline user portrait based on the processed data in the offline data; if the offline data is completely processed, updating the third offline user portrait according to the first temporary table to obtain a second real-time user portrait, and the third offline user portrait is obtained by updating the first offline user portrait based on the offline data.

[0104] Storage medium - any of various types of memory devices or storage devices. The term "storage medium" is intended to include: installation media, such as CD-ROM, floppy disk or tape device; computer system memory or random access memory, such as DRAM, DDR RAM, SRAM, EDO RAM, Rambus RAM, etc.; non-volatile memory, such as flash memory, magnetic media (such as hard disk or optical storage); registers or other similar types of memory elements, etc. Storage media may also include other types of memory or combinations thereof. In addition, the storage medium may be located in the first computer system in which the program is executed, or may be located in a different second computer system, which is connected to the first computer system via a network (such as the Internet). The second computer system can provide program instructions to the first computer for execution. The term "storage medium" may include two or more storage media residing in different locations (for example, in different computer systems connected by a network). The storage medium may store program instructions (for example, embodied as a computer program) that can be executed by one or more processors.

[0105] Of course, the storage medium containing computer executable instructions provided in the embodiment of the present application is not limited to the user portrait updating method as described above, and the computer executable instructions can also execute the related operations in the user portrait updating method provided in any embodiment of the present application.

[0106] The user portrait updating apparatus, storage medium and user portrait updating device provided in the above embodiments can execute the user portrait updating method provided in any embodiment of the present application. For technical details not described in detail in the above embodiments, please refer to the user portrait updating method provided in any embodiment of the present application.

[0107] The above are only preferred embodiments of the present application and the technical principles used. The present application is not limited to the specific embodiments herein, and various obvious changes, readjustments and substitutions that can be made by those skilled in the art will not deviate from the protection scope of the present application. Therefore, although the present application is described in more detail through the above embodiments, the present application is not limited to the above embodiments, and may include more other equivalent embodiments without departing from the concept of the present application, and the scope of the present application is determined by the scope of the claims.

Claims

1. A user portrait updating method, characterized in that: include: Acquire offline data collected within a corresponding time period at a first time node, and update a first offline user profile generated at a second time node according to the offline data, wherein the time interval between the second time node and the first time node is the time period corresponding to the first time node; the offline data includes first behavior data, first content data, first access path data, first access time data, first geographic location information, and first device information of the user; When receiving real-time data, adding the real-time data to the first temporary table and determining whether the offline data is completely processed; the real-time data includes the user's second behavior data, second content data, second access path data, second access time data, second geographic location information and second device information; If the offline data is not completely processed, a second temporary table is generated according to the unprocessed data in the offline data, and a second offline user portrait is updated according to the second temporary table and the first temporary table to obtain a first real-time user portrait, and the second offline user portrait is obtained by updating the first offline user portrait based on the processed data in the offline data; If the offline data is completely processed, the third offline user portrait is updated according to the first temporary table to obtain a second real-time user portrait, and the third offline user portrait is obtained by updating the first offline user portrait based on the offline data.

2. The user portrait updating method according to claim 1, characterized in that: The updating of the first offline user portrait generated at the second time node according to the offline data includes: Extracting features from the offline data to obtain first feature information, where the first feature information includes a first behavior feature, a first content feature, a first path feature, a first time feature, a first location feature, a first device feature, and a first statistical feature; Inputting the first feature information into a pre-trained first clustering model to obtain a first group category output by the first clustering model; A first user tag is determined according to the first feature information and the first group category, and a user tag of the first offline user portrait is updated according to the first user tag.

3. The user portrait updating method according to claim 2, characterized in that: The adding the real-time data to the first temporary table includes: Acquire second behavior data and second content data from the real-time data; The second behavior data and the second content data are associated and added to the first temporary table.

4. The user portrait updating method according to claim 2, characterized in that: The generating a second temporary table according to the unprocessed data in the offline data includes: First behavior data and corresponding first content data are obtained from the unprocessed data, and the first behavior data and the corresponding first content data are associated and added to a second temporary table.

5. The user portrait updating method according to claim 3, characterized in that: The updating the second offline user portrait according to the second temporary table and the first temporary table to obtain the first real-time user portrait includes: Extracting a second behavior feature, a second content feature, and a second statistical feature according to the second behavior data and the second content data in the first temporary table and the first behavior data and the first content data in the second temporary table; Inputting the second behavior feature, the second content feature, and the second statistical feature into a pre-trained second grouping model to obtain a second group category output by the second grouping model; A second user tag is determined according to the second group category, the second behavior feature, the second content feature and the second statistical feature, and a user tag of a second offline user portrait is updated according to the second user tag to obtain a first real-time user portrait.

6. The user portrait updating method according to claim 3, characterized in that: The updating of the third offline user portrait according to the first temporary table to obtain the second real-time user portrait includes: Extracting a third behavior feature, a third content feature, and a third statistical feature according to the second behavior data and the second content data in the first temporary table; Inputting the third behavior feature, the third content feature and the third statistical feature into a pre-trained second grouping model to obtain a third group category output by the second grouping model; A third user tag is determined according to the third group category, the third behavior feature, the third content feature and the third statistical feature, and a user tag of a third offline user portrait is updated according to the third user tag to obtain a second real-time user portrait.

7. The user portrait updating method according to claim 1, characterized in that: After the second offline user portrait is updated according to the second temporary table and the first temporary table to obtain the first real-time user portrait, the method further includes: Performing feature extraction on the unprocessed data to obtain second feature information, inputting the second feature information into a pre-trained first clustering model, and obtaining a fourth group category output by the first clustering model; A fourth user tag is determined according to the second feature information and the fourth group category, and a user tag of the second offline user portrait is updated according to the fourth user tag.

8. A user portrait updating device, characterized in that: include: A first offline update module is configured to obtain offline data collected within a corresponding time period at a first time node, and update a first offline user portrait generated at a second time node according to the offline data, wherein the time interval between the second time node and the first time node is the time period corresponding to the first time node; the offline data includes the user's first behavior data, first content data, first access path data, first access time data, first geographic location information, and first device information; an update detection module, configured to, upon receiving real-time data, add the real-time data to a first temporary table and determine whether the offline data is completely processed; the real-time data includes the user's second behavior data, second content data, second access path data, second access time data, second geographic location information, and second device information; A first real-time update module is configured to generate a second temporary table according to the unprocessed data in the offline data if the offline data is not completely processed, and update the second offline user portrait according to the second temporary table and the first temporary table to obtain a first real-time user portrait, wherein the second offline user portrait is obtained by updating the first offline user portrait based on the processed data in the offline data; The second real-time update module is configured to update the third offline user portrait according to the first temporary table to obtain a second real-time user portrait if the offline data is completely processed, and the third offline user portrait is obtained by updating the first offline user portrait based on the offline data.

9. A user portrait updating device, characterized in that: include: one or more processors; A memory stores one or more programs, and when the one or more programs are executed by the one or more processors, the one or more processors implement the user portrait updating method as described in any one of claims 1-7.

10. A storage medium containing computer executable instructions, characterized in that: The computer executable instructions, when executed by a computer processor, are used to execute the user portrait updating method as described in any one of claims 1-7.

Images