Content recommendation method and apparatus, and device

Abstract

The present application provides a content recommendation method and apparatus, and a device. The method comprises: screening a plurality of pieces of financial information to obtain a plurality of pieces of candidate information; for any candidate information among the plurality of pieces of candidate information, generating an information summary card of the candidate information and a card tag of the information summary card on the basis of information content of the candidate information; acquiring target user profile data of a target user; on the basis of the target user profile data and the plurality of card tags, screening the plurality of information summary cards to obtain at least one content card to be recommended; and pushing the at least one content card to be recommended to an application client of the target user. Therefore, the content recommendation accuracy is improved, and the content recommendation effect is improved.

Description

Content recommendation method, device and equipment

[0001] Priority information

[0002] This application claims priority to Chinese patent application number "202311655884.X" filed on December 1, 2023, entitled "Content recommendation method, device and equipment", and the entire contents of which are incorporated by reference into this application. Technical Field

[0003] The embodiments of the present application relate to the field of computer technology, and in particular to a content recommendation method, apparatus, and device. Background Art

[0004] Currently, when a client recommends information to a user, the title of the information is generally displayed in the information list. The user needs to click on the displayed information title and then enter the consultation details page to browse the full text of the information.

[0005] However, due to the large amount of information recommended by the client, users need to spend a lot of time reading a large amount of information and need to summarize the information themselves before they can get the effective information they need. Obviously, this recommendation method has poor recommendation effect and low recommendation accuracy. Summary of the Invention

[0006] The present application provides a content recommendation method, apparatus, and device to improve the accuracy of content recommendation and enhance the effect of content recommendation.

[0007] In a first aspect, the present application provides a content recommendation method, including: screening multiple financial information to obtain multiple candidate information; for any candidate information among the multiple candidate information, generating an information summary card and a card label of the information summary card based on the information content of the candidate information; obtaining target user portrait data of the target user; screening multiple information summary cards based on the target user portrait data and multiple card labels to obtain at least one content card to be recommended; and pushing at least one content card to be recommended to the application client of the target user.

[0008] In the second aspect, the present application provides a content recommendation device, including a first screening module for screening multiple financial information to obtain multiple candidate information; a first generation module for generating an information summary card and a card label of the information summary card for any candidate information among the multiple candidate information based on the information content of the candidate information; a first acquisition module for obtaining target user portrait data of the target user; a second screening module for screening multiple information summary cards based on the target user portrait data and multiple card labels to obtain at least one content card to be recommended; and a push module for pushing at least one content card to be recommended to the application client of the target user.

[0009] In a third aspect, the present application provides an electronic device comprising: a processor and a memory, the memory being used to store a computer program, the processor being used to call and run the computer program stored in the memory to execute the method as in the first aspect or its various implementations.

[0010] In a fourth aspect, the present application provides a computer-readable storage medium for storing a computer program, wherein the computer program enables a computer to execute the method as in the first aspect or its various implementations.

[0011] In a fifth aspect, the present application provides a computer program product, comprising computer program instructions, which enable a computer to execute the method in the first aspect or its various implementations.

[0012] In a sixth aspect, the present application provides a computer program, which enables a computer to execute the method in the first aspect or its various implementations.

[0013] Through the technical solution of the present application, the electronic device can screen multiple financial information to obtain multiple candidate information. Then, for any candidate information among the multiple candidate information, an information summary card of the candidate information and a card label of the information summary card can be generated according to the information content of the candidate information. Then, the electronic device can obtain the target user portrait data of the target user, and screen the multiple information summary cards according to the target user portrait data and multiple card labels to obtain at least one content card to be recommended; and push the at least one content card to be recommended to the application client of the target user. In the above process, the electronic device can preliminarily screen the information to determine the candidate information, and screen the information summary cards according to the user portrait data and the card labels to determine the content card to be recommended, and push the content card to be recommended to the client. In this way, not only can the information that the user is most likely to need be selected from the massive amount of information, but also the user can understand the main content of the information in a timely and quick manner, thereby helping the user to more accurately obtain information that is relevant to themselves and helpful for investment, and facilitate efficient interpretation of information. Therefore, the accuracy of content recommendation can be improved and the effect of content recommendation can be improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0015] FIG1 is a diagram of an application scenario provided by an embodiment of the present application;

[0016] FIG2 is a flow chart of a content recommendation method provided in an embodiment of the present application;

[0017] FIG3 is a schematic diagram of a content recommendation method provided in an embodiment of the present application;

[0018] FIG4 is a schematic diagram of another content recommendation method provided in an embodiment of the present application;

[0019] FIG5 is a schematic diagram of another content recommendation method provided in an embodiment of the present application;

[0020] FIG6 is a schematic diagram of another content recommendation method provided in an embodiment of the present application;

[0021] FIG7 is a schematic diagram of another content recommendation method provided in an embodiment of the present application;

[0022] FIG8 is a schematic diagram of another content recommendation method provided in an embodiment of the present application;

[0023] FIG9 is a schematic diagram of another content recommendation method provided in an embodiment of the present application;

[0024] FIG10 is a schematic diagram of a content recommendation device 1000 provided in an embodiment of the present application;

[0025] FIG11 is a schematic block diagram of an electronic device 1100 provided in an embodiment of the present application. DETAILED DESCRIPTION

[0026] The following will be combined with the accompanying drawings in the embodiments of this application to clearly and completely describe the technical solutions in the embodiments of this application. Obviously, the embodiments described are only part of the embodiments of this application, not all of the embodiments. Based on the embodiments of this application, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of this application.

[0027] The following will be combined with the accompanying drawings in the embodiments of this application to clearly and completely describe the technical solutions in the embodiments of this application. Obviously, the embodiments described are only part of the embodiments of this application, not all of the embodiments. Based on the embodiments of this application, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of this application.

[0028] It should be noted that the terms "first", "second", etc. in the specification and claims of this application and the above-mentioned drawings are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments of the application described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions. For example, a process, method, system, product, or server that includes a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products, or devices.

[0029] As mentioned above, in the current recommendation method, users need to spend a lot of time reading a large amount of information and need to summarize the information by themselves in order to obtain the effective information they need. This has the problems of poor recommendation effect and low recommendation accuracy.

[0030] In order to solve the above technical problems, this application can perform preliminary screening of information to determine candidate information, and screen information summary cards based on user portrait data and card labels to determine the content cards to be recommended, and finally push the content cards to be recommended to the client. This can not only select the information that users are most likely to need from massive information, but also enable users to understand the main content of the information in a timely and quick manner, thereby helping users to more accurately obtain information that is relevant to themselves and helpful for investment, and facilitate efficient interpretation of information. Therefore, it can improve the accuracy of content recommendations and enhance the content recommendation effect.

[0031] It should be understood that the technical solution of this application can be applied to the following scenarios, but is not limited to:

[0032] It should be noted that the solution provided in the embodiments of the present application can be executed by any electronic device with content recommendation capabilities, and the present application does not impose any specific restrictions on the electronic device. For example, the electronic device can be a server. The server can be a single server, a server cluster composed of multiple servers, or a cloud platform control center, but is not limited thereto. Alternatively, the electronic device can be a terminal device. The terminal device can be a mobile phone, a tablet computer, a laptop computer, or a desktop computer, etc. Alternatively, the electronic device can be implemented as a combination of a server and a terminal device, wherein the server and the terminal device can communicate wirelessly or wired.

[0033] In some implementations, as shown in FIG1 , the application scenario may include a terminal device 110 and a server 120 . The terminal device 110 may be connected to the server 120 via a wired network or a wireless network. The terminal device may be the aforementioned terminal device.

[0034] For example, an application client may be installed on the terminal device 110, and the server 120 may determine multiple financial information and filter the financial information to obtain content cards to be recommended. The server 120 may then send the content cards to the terminal device 110, so that the terminal device 110 can display the content cards.

[0035] After introducing the application scenarios of the embodiments of the present application, the technical solutions of the present application will be described in detail below:

[0036] Figure 2 is a flow chart of a content recommendation method provided by an embodiment of the present application. The method can be performed by an electronic device in the above application scenario, but is not limited thereto. As shown in Figure 2, the method may include the following steps:

[0037] S210: Screening multiple financial information to obtain multiple candidate information;

[0038] S220: For any candidate information among the plurality of candidate information, generating an information summary card and a card label for the information summary card based on the information content of the candidate information;

[0039] S230: Obtain target user portrait data of the target user;

[0040] S240: Filtering multiple information summary cards based on the target user portrait data and multiple card tags to obtain at least one content card to be recommended;

[0041] S250: Push at least one content card to be recommended to the application client of the target user.

[0042] It should be noted that, in the following embodiments, this application will introduce the technical solution of this application by taking the electronic device as an example of a server. When the technical solution of this application is executed by other electronic devices, the corresponding content is similar to that here. To avoid repetition, this application will not elaborate here.

[0043] In some possible implementations, the above-mentioned multiple financial information can be any information released by any user based on a terminal device, and this application does not limit the source of the financial information.

[0044] In some implementations, the virtual resources in this application may be stocks or funds, but are not limited thereto.

[0045] In some implementable embodiments, the above-mentioned screening of multiple financial information to obtain multiple candidate information may include: determining the financial information that meets preset conditions among the multiple financial information as multiple candidate information; wherein the preset conditions include at least one of the following: the number of words in the information belongs to the word number range, the information category belongs to the preset category, there are related virtual resources, there are opinion tendencies, the timeliness of the information belongs to the timeliness range, the review status is passed, the information display status is available for external display, and the information visibility area includes the area of ​​the target user.

[0046] Among them, information category refers to the classification results of financial information in different dimensions; for example, the information category can be the content type corresponding to the financial information, such as pure text type, graphic type, and video type, etc.; for another example, the information category can be the source type corresponding to the financial information, such as the announcement type released by the official certified account, the research report type released by the third-party organization, and the discussion article type released by the ordinary account, etc.; the preset category can be the category that the target user likes or has browsed in different dimensions, but is not limited to this; for example, the preset category can be the pure text type and the research report type.

[0047] An associated virtual resource may refer to a virtual resource related to the subject or content described in the information. For example, if financial information 1 is about the price changes of stock 1 over the past year, then it can be determined that financial information 1 involves stock 1, and the virtual resource associated with financial information 1 is determined to be stock 1. In other words, it can be determined that financial information 1 has an associated virtual resource.

[0048] The presence of a biased viewpoint means that the information disclosed contains only a single opinion or conclusion. For example, if Financial News 2 only describes that Stock 1 has been on an upward trend over the past week, then it can be determined that Financial News 2 contains only one opinion: "Stock 1 has been on an upward trend over the past week." In other words, it can be determined that Financial News 2 has a biased viewpoint. If Financial News 3 only discloses listed companies in Region 1, then it can be determined that Financial News 3 does not disclose any opinions or conclusions, meaning that it has no biased viewpoint.

[0049] Information timeliness can refer to the time during which the information can be made public. For example, assuming that financial information 4 is set to be visible for only 3 days and the release time of financial information 4 is November 10, then the information timeliness of financial information 4 can be determined to be from November 10 to November 12. Information timeliness can also refer to the validity period of the content disclosed by the information. For example, assuming that financial information 5 discloses that "you can participate in the lottery on November 13", then the validity period of financial information 5 can be determined to be on or before November 13. Information timeliness can also refer to the release time of the information. For example, the timeliness interval can be set to 48 hours. When the release time of the information is within 48 hours of the current time, it can be determined that the information timeliness falls within the timeliness interval. When the release time of the information is before 48 hours of the current time, it can be determined that the information timeliness does not fall within the timeliness interval.

[0050] The review status can include unreviewed, under review, and approved. For example, after financial information is released, the application client staff can first review it and determine and record its review status based on the review progress. The server can then query the review status of the recorded financial information.

[0051] The information display status also refers to the visibility status of the information, which can include being visible to yourself and being open to the public. The information display status of "visible only to yourself" means that the information can only be displayed on the client corresponding to the publisher of the information. The information display status of "open to the public" means that the information can be displayed on any client.

[0052] The target user's region may be the region of the communication network used by the target user's client, or the region set by the target user based on the client, but is not limited thereto. The information visibility region may be the region set by the information publisher based on the client.

[0053] Exemplarily, the server may perform a test on the financial information to determine whether the financial information meets preset conditions. For example, the server may perform a character count test on the financial information to determine the word count, and then determine whether the determined word count falls within a word count range. The server may also perform a content test on the financial information to determine whether the financial information includes a virtual resource identifier. The virtual resource identifier may be a name or code of a virtual resource. If the virtual resource name is included, it can be determined that an associated virtual resource exists; if the virtual resource name is not included, it can be determined that an associated virtual resource does not exist.

[0054] In the above process, the electronic device can first perform an initial screening of the financial information to ensure that the information recommended to the user meets the preset conditions. For example, it can ensure that the recommended information is timely and the categories are what the user likes or pays attention to, etc., thereby ensuring the quality of the recommended information.

[0055] In some implementations, the server may extract key content from the candidate information using an artificial intelligence-based algorithm or model, thereby determining an information summary card and a card label for the information summary card.

[0056] For example, the server can use the TextRank algorithm to extract information from the title and content of the candidate information, thereby obtaining summary information and card labels. Then, the summary information can be processed based on the card flow technology to obtain information summary cards, wherein the information summary card can include summary information and, of course, the title of the candidate information. In this way, the target user's application client can display information in the form of cards, so that the user can read the summary information of the information more intuitively and interpret the information efficiently, thereby avoiding the need for users to read massive amounts of information before understanding the function of the information and whether the information is beneficial to their investment decisions. It allows users to quickly understand and interpret the core information involved in the information, efficiently obtain and process information that is helpful for investment, discover more investment opportunities, and assist in investment decisions.

[0057] Card labels may include, but are not limited to, the candidate information's information identifier (IDentity), public opinion label, content language, regional visibility, total number of words in the information, information category, and information release time. Content language refers to the language used when writing the candidate information; regional visibility refers to the regions where the candidate information can be displayed; public opinion labels may represent the opinion of the candidate information, for example, whether the content of the candidate information is bearish or bullish on the associated stock; total number of words in the information refers to the total number of words in the information summary card corresponding to the original candidate information; and information category refers to the classification results of the original candidate information in different dimensions corresponding to the information summary card.

[0058] Among them, user portrait data refers to the user information profile or user feature description formed by collecting, integrating and analyzing multi-dimensional data such as user behavior; specifically, user portrait data may include basic information of the user's corresponding client account, such as account name, account ID, client login region, client language setting, etc., and may also include behavioral information of the user's account, such as the information read and related information, such as reading time, the market corresponding to the information, the language setting corresponding to the information, etc., virtual resources corresponding to financial-related transactions and related information, such as the category of virtual resources traded, transaction frequency, transaction amount, etc.

[0059] In some implementable embodiments, the above-mentioned screening of multiple information summary cards according to the target user portrait data and multiple card tags to obtain at least one content card to be recommended further includes: determining the virtual resource identifier of the virtual resource associated with the candidate information according to the information content of the candidate information; determining the mapping relationship between the virtual resource identifier and the information summary card of the candidate information; the above-mentioned screening of multiple information summary cards according to the target user portrait data and multiple card tags to obtain at least one content card to be recommended includes: obtaining the target virtual resource identifier of the target virtual resource according to the target user portrait data, wherein the target virtual resource includes the virtual resource associated with the target user; determining at least one initial content card from the multiple information summary cards through the mapping relationship according to the target virtual resource identifier; and determining at least one content card to be recommended from the at least one initial content card according to the card tag of the at least one initial content card.

[0060] Exemplarily, the virtual resource associated with the candidate information may be a virtual resource included in the title or content of the candidate information. The server may identify a target keyword in the title or content of the candidate information. The target keyword may be stocks, funds, etc., and then determine the words including the target keyword as the virtual resource associated with the candidate information. For example, assuming the target keyword is stocks, and the content of financial information 1 includes: the price changes of stock 1 in the past year, the server may identify the target keyword "stocks" and extract the words including the target keyword "stocks" as "stock 1", thereby determining that the virtual resource associated with financial information 1 is stock 1.

[0061] Exemplarily, the identifier of the virtual resource may be used to uniquely identify the virtual resource, such as the name or stock code of the virtual resource, but is not limited thereto.

[0062] Specifically, by identifying the information content of the candidate information and determining the virtual resource identifier of the virtual resource associated with each candidate information, the virtual resource identifier of the virtual resource associated with the information summary card corresponding to each candidate information can be obtained; then, through reverse calculation, the information summary cards associated with different virtual resource identifiers (i.e., different virtual resources) are obtained, and a corresponding mapping relationship is established, which is convenient for subsequent virtual resource identifiers based on the virtual resources associated with the target user. Through this mapping relationship, the mapped information summary card can be quickly found to obtain the initial content card, thereby improving the screening accuracy and efficiency of the initial content card.

[0063] Exemplarily, the target virtual resource includes at least one of the following: a virtual resource that the target user is interested in, and a virtual resource involved in the target user's historical browsing content.

[0064] Exemplarily, when determining at least one initial content card from a plurality of information summary cards through a mapping relationship according to the target virtual resource identifier, the information summary card having a mapping relationship with the target virtual resource identifier may be determined as the initial content card.

[0065] Exemplarily, before obtaining the target user portrait data of the target user, the above-mentioned process also includes: obtaining the virtual resources that the user is concerned about and stored in the data center to obtain virtual resource list data; storing the virtual resource list data in a public database; correspondingly, obtaining the target virtual resource identifier of the target virtual resource based on the target user portrait data, including: obtaining the target virtual resource that the target user is concerned about from the virtual resource list data in the public database based on the user identifier in the target user portrait data; and determining the target virtual resource identifier of the target virtual resource.

[0066] For example, the virtual resource list data may include a user ID and the virtual resources that the user is interested in corresponding to the user ID. The server may search for the corresponding virtual resource from a public database based on the user ID in the target user portrait data and determine the virtual resource as the target virtual resource.

[0067] It is understandable that when a user follows a certain stock based on an application terminal, the terminal device can store the data corresponding to this user operation in the data middle platform. After that, the server can obtain the stock that the user follows from the data middle platform, thereby determining the target virtual resource. However, since there is a large delay when the server obtains the virtual resource corresponding to the target user from the data middle platform, that is, the real-time performance is poor, and in other scenarios, other services or servers may also use the data of the virtual resource that the user follows. Therefore, as shown in Figure 3, the server can create a public user feature service. The public user feature service includes a data task. By executing the data task, the data of the virtual resource that the user follows in Kafka, that is, the virtual resource list data, is written into a public database such as a redis database. After that, the server can obtain the virtual resource that the target user follows from the redis database based on the public user feature service, thereby reducing the delay in obtaining the target virtual resource and improving the real-time performance of data acquisition. Among them, Kafka is a message middleware. When a user follows a virtual resource based on a terminal device, the terminal device can report the data corresponding to the virtual resource to kafka.

[0068] Exemplarily, the above-mentioned determining at least one content card to be recommended from at least one initial content card based on the card tag of at least one initial content card includes: determining filtering rule data associated with the target user based on the target user portrait data, wherein the filtering rule data includes at least one of the following: historical recommended content, language preference information, regional information, and information browsing time information; filtering at least one initial content card based on the filtering rule data and the card tag of at least one initial content card to determine at least one content card to be recommended.

[0069] For example, information browsing time information can be the average time the target user has spent browsing information in the past. This information can be used to measure whether the target user prefers to read short or long articles, thereby improving the accuracy of subsequent news summary card recommendations to the user. Regional information can include the financial market corresponding to the information previously browsed by the target user, such as information on the Chinese or Singaporean financial markets, or the location of the application client, such as the United States, China, or Singapore. Language preference information can be the system language of the application client. Historically recommended content can be the card tags of information previously recommended to the application client. After determining the filtering rule data associated with the target user, at least one initial content card can be initially screened based on the historically recommended content. Initial content cards recommended to the target user's client can be filtered to obtain first content cards that the target user has not read. Furthermore, the target user's language preference information, regional information, and information browsing time information are matched with the card tags corresponding to each of the remaining first content cards for further screening, thereby obtaining information summary cards to be recommended to the target user. Specifically, the language preference information and regional information of the target user can be matched with the content language and regional visibility status in the corresponding card labels of the remaining first content cards, so as to filter out the second content card with the same language and visible to the corresponding regional information of the target user from the first content card, and then determine the target user's recommended reading word count based on the target user's information browsing time information, and compare the full word count of the information in the corresponding card label of the second content card with the recommended reading word count of the target user, so as to filter out the third content card with the full word count of the information less than the recommended reading word count from the second content card as the final content card to be recommended.

[0070] For example, when filtering initial content cards according to filtering rule data, the similarity between the card tags and the filtering rule data may be calculated, and initial content cards with similarity greater than a similarity threshold may be determined as content cards to be recommended.

[0071] In some implementable embodiments, before pushing at least one content card to be recommended to the target user's application client, the content card to be recommended can be specifically sorted according to the information category in the card label corresponding to the content card to be recommended (for example, it can be sorted according to the priority of the information category) to filter out the recommended content cards with a preset number of rankings at the front as the information summary cards inserted into the above-mentioned community recommendation flow. For example, the information category refers to the source type corresponding to the financial information, including the announcement type from the official certified account, the research report type from the third-party agency, and the discussion article type from the ordinary account; when sorting the recommended content cards, the priority of the announcement type is greater than the priority of the research report type, and the priority of the research report type is greater than the priority of the discussion article type. Specifically, the content card to be recommended can be sorted according to the information category in the card label corresponding to the content card to be recommended to filter out the recommended content cards with a preset number of rankings (such as a preset number of 1) as the information summary cards inserted into the above-mentioned community recommendation flow.

[0072] In some implementable embodiments, before pushing at least one to-be-recommended content card to the target user's application client, the historical users corresponding to the to-be-recommended content card may be obtained, where the historical users refer to users who have clicked on or viewed the to-be-recommended content card, and user profile data of the historical users may be clustered to obtain a cluster center as the target user feature data corresponding to the to-be-recommended content card. Furthermore, based on the user profile data of the target user and the target user feature data, the recommendation degree for recommending the to-be-recommended content card to the application client is determined. Finally, based on the recommendation degree corresponding to each to-be-recommended content card, each to-be-recommended content card is sorted to select a preset number of to-be-recommended content cards ranked first as the information summary cards inserted into the community recommendation flow. Specifically, the recommendation degree for recommending the to-be-recommended content card to the application client may be calculated using the following formula:

[0073] Among them, feature_op={x1,…,x n} represents the target user feature data corresponding to the content card to be recommended, x i (i=1,2,…,n) represents the i-th feature information in the target user feature data; feature_cur={y1,…,y n} represents the user portrait data of the target user, y i (i=1,2,…,n) represents the i-th feature information in the user portrait data of the target user.

[0074] It is understandable that sim(feature op ,feature cur) is used to indicate the distance between the target user's characteristic data and the target user's user portrait data, that is, it can indicate the degree of similarity between the historical user corresponding to the content card to be recommended and the target user. In other words, the smaller the distance between the target user's characteristic data and the target user's user portrait data, the more similar the historical user corresponding to the content card to be recommended is to the target user, and the higher the ranking of the content card to be recommended is.

[0075] In some implementations, before pushing the at least one content card to the target user's application client, the method may further include: determining a display order for each of the at least one content card based on display attribute information of the at least one content card, where the display attribute information includes at least one of the following: information release time, information query word count, number of information likes, number of information forwards, and the first letter of the information title. Pushing the at least one content card to the target user's application client includes: pushing the at least one content card to the application client based on the display order of the at least one content card.

[0076] Exemplarily, the server may first quantify the display attribute information to determine the numerical value corresponding to each display attribute information. Then, for any content card to be recommended, the server may calculate the weighted sum of at least one display attribute information of the content card to be recommended. Then, the server may sort the weighted sum corresponding to each content card to be recommended to obtain the display order of each content card to be recommended.

[0077] In the above process, the server can further filter the information summary cards according to the filtering rule data and card labels, so as to ensure that users can obtain information related to them more accurately, improve the accuracy of content recommendation, and enhance the content recommendation effect.

[0078] In some implementations, pushing at least one content card to be recommended to the target user's application client includes: determining an information recommendation feed stream and a target insertion location; adding at least one content card to be recommended at the target insertion location in the information recommendation feed stream; and pushing the added information recommendation feed stream to the application client. Both the card stream and the feed stream are information streams that present content.

[0079] For example, as shown in Figure 4, the community recommendation stream can refer to an existing feed stream for recommending information. Based on the user profile service, the server can retrieve the ID of the stock that the user is interested in, and then recall the information summary card corresponding to the stock ID. This means determining the initial content card related to the stock corresponding to the stock ID. The initial content card is then filtered to obtain the content cards to be recommended. After sorting the content cards to be recommended, the content cards to be recommended can be inserted into a fixed position in the community recommendation stream and pushed to the application client.

[0080] The present application will further introduce the technical solution of the present application in the form of schematic diagrams in the following embodiments:

[0081] For example, as shown in Figure 5, the server may include an information service, an algorithm service, a card generation service, and a community backend. The card generation service may include an artificial intelligence model for extracting key content. After different users publish financial information through the client, the information service may determine that new financial information has been generated and notify the algorithm service. The algorithm service, through a series of judgment logic, may filter out information that meets the recommendation requirements, obtain candidate information, and use the card generation service to generate information summaries based on the candidate information. The algorithm service then generates information summary cards based on the community backend based on the information summaries. Subsequently, as shown in Figure 6, when the user opens the client, the client can pull recommendation data from the community backend. At this time, the community backend will request a personalized recommendation list. As part of the community content pool, the information summary cards can be distributed through the community's personalized recommendation list. In other words, the community backend can return information summary cards to the client, but the community backend can determine specific information summary cards through the algorithm service and return them to the client. Specifically, the community backend can first pull the latest recommendation data, such as information summary cards, from the algorithm side service. Then, the algorithm side service can use the user portrait data based on the method in the above embodiment to perform logical processing such as recalling, filtering, and sorting on the information summary cards in the above-generated recommendation data to obtain the content card to be recommended, that is, the specific information summary card, and insert the specific information summary card into the existing recommendation data stream and return it to the community backend. The community backend can perform business logic processing on it and return it to the client for display.

[0082] Exemplarily, the server can determine the corresponding information summary card from multiple information summary cards and recommend it to the user based on the user's characteristics and the stocks that the user cares about. Specifically, as shown in Figure 7, the server can obtain user characteristics, such as user portrait data, and recall relevant cards based on the list of stocks that the user cares about and / or the list of stocks that have been recently browsed. Recalling relevant cards can refer to using the stock ID in the list of stocks that the user cares about to match the IDs of the stocks associated with multiple information summary cards. The information summary card that successfully matches is the relevant card that needs to be recalled, that is, the initial content card. Afterwards, the card characteristics of the card, that is, the card label, can be obtained, and the above-mentioned related cards can be filtered and sorted according to the user portrait and card label. Afterwards, the sorted cards can be inserted into the original feed stream of recommended information and recommended to the client.

[0083] Exemplarily, in combination with the above embodiment, as shown in FIG8 , the server may include a consumption task module, a timed task module, a community card generation interface, and a public opinion and card summary interface. The consumption task module may obtain multiple financial information from Kafka, and perform a first round of filtering, i.e., pre-filtering, on the multiple financial information based on the visible area of ​​the financial information, whether the financial information is open to users, the title of the information, the content of the information body, the writing language of the financial information, the jump link of the financial information, the timeliness of the financial information, the category of the financial information, the number of words in the financial information, whether the financial information involves virtual resources, the review status of the financial information, etc., to obtain a candidate information set (information candidate table). The candidate information set may include multiple candidate information, and the information is maintained in MySQL. Next, the timed task module may periodically pull the information set for which no card has been generated from the information candidate set, and input the title and content of the information into the public opinion and card summary interface to obtain the public opinion label and the card summary, i.e., summary information. The public opinion label may represent the opinion tendency of the information. Next, for news items that can generate cards, the scheduled task can call the community card generation API to generate a news summary card and card tags based on the news item's ID, public opinion tag, summary information, news writing language, publishable regions, whether the news is publicly available, news title, and the information's jump link. The scheduled task module can then send the news summary card, or a consultation summary card filtered by card tags, to the community feed generation API in the community backend. Finally, the scheduled task module can update the MySQL table with three fields: public opinion tag, card summary, and whether a card has been generated.

[0084] In some possible implementations, in combination with the above embodiment, as shown in FIG9 , the server may first obtain multiple information summary cards, determine the associated information summary cards involving virtual resources (i.e., related to virtual resources), and then filter the associated information summary cards according to the information type, whether the information can be recommended, and the timeliness of the information, and then determine the stock IDs involved in the filtered information summary cards. Next, the mapping relationship between the stock ID and the filtered information summary card can be determined by an inverted algorithm, and the mapping relationship can be stored in the Redis database. When recalling, that is, when recommending information to the user again, the corresponding information summary card can be directly obtained from the mapping relationship according to the stock ID of the stock that the user is following, and recommended to the client where the user is logged in, thereby improving the efficiency of card recommendation.

[0085] It should be noted that all the above technical solutions can be combined in any way to form optional embodiments of the present application, and will not be described in detail here.

[0086] It should be noted that in the specific implementation of this application, when information, user portrait data, virtual resources that are followed or browsed, and other related data are involved, when the embodiments of this application are applied to specific products or technologies, user permission, consent or authorization must be obtained, and the collection, use and processing of relevant data must comply with relevant laws, regulations and standards of relevant countries and regions.

[0087] Figure 10 is a schematic diagram of a content recommendation device 1000 provided in an embodiment of the present application. As shown in Figure 10 , content recommendation device 1000 includes: a first screening module 1001, a first generation module 1002, a first acquisition module 1003, a second screening module 1004, a push module 1005, a first determination module 1006, a second determination module 1007, a second acquisition module 1008, a storage module 1009, and a third determination module 1010.

[0088] In some possible implementations, the first screening module 1001 is used to screen multiple financial information to obtain multiple candidate information; the first generation module 1002 is used to generate an information summary card and a card label of the information summary card for any candidate information among the multiple candidate information based on the information content of the candidate information; the first acquisition module 1003 is used to obtain the target user portrait data of the target user; the second screening module 1004 is used to screen multiple information summary cards based on the target user portrait data and multiple card labels to obtain at least one content card to be recommended; the push module 1005 is used to push at least one content card to be recommended to the application client of the target user.

[0089] In some possible implementations, the first determination module 1006 is used to determine the virtual resource identifier of the virtual resource associated with the candidate information based on the information content of the candidate information; the second determination module 1007 is used to determine the mapping relationship between the virtual resource identifier and the information summary card of the candidate information; the second screening module 1004 is specifically used to: obtain the target virtual resource identifier of the target virtual resource based on the target user portrait data, wherein the target virtual resource includes the virtual resource associated with the target user; determine at least one initial content card from multiple information summary cards through the mapping relationship based on the target virtual resource identifier; and determine at least one content card to be recommended from at least one initial content card based on the card label of the at least one initial content card.

[0090] In some possible implementations, the second screening module 1004 is specifically used to: determine filtering rule data associated with the target user based on the target user portrait data, wherein the filtering rule data includes at least one of the following: historical recommended content, language preference information, regional information, and information browsing time information; filter at least one initial content card based on the filtering rule data and the card label of at least one initial content card to determine at least one content card to be recommended.

[0091] In some possible implementations, the second acquisition module 1008 is used to acquire the virtual resources that the user is interested in and stored in the data center to obtain virtual resource list data; the storage module 1009 is used to store the virtual resource list data in a public database; the second screening module 1004 is specifically used to: acquire the target virtual resources that the target user is interested in from the virtual resource list data in the public database based on the user identifier in the target user portrait data; and determine the target virtual resource identifier of the target virtual resource.

[0092] In some possible implementations, the first screening module 1001 is specifically used to: determine financial information that meets preset conditions among multiple financial information as multiple candidate information; wherein the preset conditions include at least one of the following: the number of words in the information belongs to the word number range, the information category belongs to the preset category, there are related virtual resources, there are opinion tendencies, the timeliness of the information belongs to the timeliness range, the review status is passed, the information display status is available for external display, and the information visibility area includes the area of ​​the target user.

[0093] In some implementations, the push module 1005 is specifically used to: determine the information recommendation feed stream and the target insertion position; add at least one content card to be recommended at the target insertion position of the information recommendation feed stream, and push the added information recommendation feed stream to the application client.

[0094] In some possible implementations, the third determination module 1010 is used to determine the display order of at least one content card to be recommended based on the display attribute information of at least one content card to be recommended, where the display attribute information includes at least one of the following: information release time, number of words in the information consultation, number of likes for the information, number of forwardings for the information, and the first letter of the title of the information; the push module 1005 is specifically used to push at least one content card to be recommended to the application client according to the display order of at least one content card to be recommended.

[0095] It should be understood that the device embodiments and method embodiments may correspond to each other, and similar descriptions may refer to the method embodiments. To avoid repetition, they will not be described in detail here. Specifically, the device 1000 shown in Figure 10 can perform the above-mentioned method embodiments, and the aforementioned and other operations and / or functions of the various modules in the device 1000 are respectively for implementing the corresponding processes in the above-mentioned various methods. For the sake of brevity, they will not be described in detail here.

[0096] The above describes the device 1000 of the embodiment of the present application from the perspective of functional modules in conjunction with the accompanying drawings. It should be understood that the functional module can be implemented in hardware form, can be implemented by instructions in software form, and can also be implemented by a combination of hardware and software modules. Specifically, the steps of the method embodiment in the embodiment of the present application can be completed by the hardware integrated logic circuit and / or software form instructions in the processor, and the steps of the method disclosed in the embodiment of the present application can be directly embodied as being executed by a hardware decoding processor, or can be executed by a combination of hardware and software modules in the decoding processor. Optionally, the software module can be located in a mature storage medium in the art such as random access memory, flash memory, read-only memory, programmable read-only memory, electrically erasable programmable memory, registers, etc. The storage medium is located in the memory, and the processor reads the information in the memory and completes the steps in the above method embodiment in conjunction with its hardware.

[0097] FIG11 is a schematic block diagram of an electronic device 1100 provided in an embodiment of the present application.

[0098] As shown in FIG11 , the electronic device 1100 may include:

[0099] The memory 1110 and the processor 1120 are configured to store computer programs and transmit the program code to the processor 1120. In other words, the processor 1120 can call and execute the computer program from the memory 1110 to implement the method in the embodiment of the present application.

[0100] For example, the processor 1120 may be configured to execute the above method embodiments according to instructions in the computer program.

[0101] In some embodiments of the present application, the processor 1120 may include but is not limited to:

[0102] General-purpose processor, Digital Signal Processor (DSP), Application Specific Integrated Circuit (ASIC), Field Programmable Gate Array (FPGA) or other programmable logic device, discrete gate or transistor logic device, discrete hardware components, etc.

[0103] In some embodiments of the present application, the memory 1110 includes but is not limited to:

[0104] Volatile memory and / or non-volatile memory. Non-volatile memory can be read-only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), or flash memory. Volatile memory can be random access memory (RAM), which is used as an external cache. By way of example and not limitation, many forms of RAM are available, such as static random access memory (SRAM), dynamic random access memory (DRAM), synchronous dynamic random access memory (SDRAM), double data rate synchronous dynamic random access memory (DDR SDRAM), enhanced synchronous dynamic random access memory (ESDRAM), synchronous link DRAM (SLDRAM), and direct rambus RAM (DR RAM).

[0105] In some embodiments of the present application, the computer program may be divided into one or more modules, which are stored in the memory 1110 and executed by the processor 1120 to implement the method provided by the present application. The one or more modules may be a series of computer program instruction segments capable of implementing specific functions, and the instruction segments are used to describe the execution process of the computer program in the electronic device.

[0106] As shown in FIG11 , the electronic device may further include:

[0107] The transceiver 1130 may be connected to the processor 1120 or the memory 1110 .

[0108] The processor 1120 may control the transceiver 1130 to communicate with other devices. Specifically, it may send information or data to other devices or receive information or data sent by other devices. The transceiver 1130 may include a transmitter and a receiver. The transceiver 1130 may further include an antenna, which may be one or more.

[0109] It should be understood that the various components in the electronic device are connected via a bus system, wherein the bus system includes not only a data bus but also a power bus, a control bus and a status signal bus.

[0110] The present application also provides a computer storage medium having a computer program stored thereon, which, when executed by a computer, enables the computer to perform the method of the above-mentioned method embodiment. In other words, the present application also provides a computer program product containing instructions, which, when executed by a computer, enables the computer to perform the method of the above-mentioned method embodiment.

[0111] When software is used to implement, it can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions. When the computer program instruction is loaded and executed on a computer, the process or function according to the embodiment of the present application is generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network or other programmable devices. The computer instruction can be stored in a computer-readable storage medium, or transmitted from a computer-readable storage medium to another computer-readable storage medium. For example, the computer instruction can be transmitted from a website, a computer, a server or a data center by wired (such as coaxial cable, optical fiber, digital subscriber line (Digital Subscriber Line, DSL)) or wireless (such as infrared, wireless, microwave, etc.) mode to another website, a computer, a server or a data center. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server, a data center that includes one or more available media integrations. The available medium can be a magnetic medium (such as a floppy disk, a hard disk, a magnetic tape), an optical medium (such as a digital video disc (Digital Video Disc, DVD)) or a semiconductor medium (such as a solid-state drive (SSD)) etc.

[0112] Those skilled in the art will appreciate that the modules and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0113] In the several embodiments provided in this application, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. For example, the device embodiments described above are merely schematic. For example, the division of the modules is merely a logical function division. In actual implementation, there may be other division methods, such as multiple modules or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or modules, which can be electrical, mechanical or other forms.

[0114] Modules described as separate components may or may not be physically separate, and components displayed as modules may or may not be physical modules, i.e., they may be located in one place or distributed across multiple network elements. Some or all of the modules may be selected based on actual needs to achieve the purpose of the present embodiment. For example, the functional modules in the various embodiments of the present application may be integrated into a processing module, or each module may exist physically separately, or two or more modules may be integrated into a single module.

[0115] The above are only specific embodiments of the present application, but the scope of protection of this application is not limited thereto. Any changes or substitutions that can be easily conceived by a person skilled in the art within the technical scope disclosed in this application should be included in the scope of protection of this application. Therefore, the scope of protection of this application should be based on the scope of protection of the claims.

Claims

1. A content recommendation method, characterized in that: include: Screen multiple financial information to obtain multiple candidate information; For any candidate information among the plurality of candidate information, generating an information summary card of the candidate information and a card label of the information summary card according to the information content of the candidate information; Obtain target user portrait data of target users; Filtering the plurality of information summary cards according to the target user portrait data and the plurality of card tags to obtain at least one content card to be recommended; The at least one content card to be recommended is pushed to the application client of the target user.

2. The method according to claim 1, characterized in that Before screening the plurality of information summary cards according to the target user portrait data and the plurality of card tags to obtain at least one content card to be recommended, the method further includes: Determining, according to the information content of the candidate information, a virtual resource identifier of a virtual resource associated with the candidate information; Determine a mapping relationship between the virtual resource identifier and the information summary card of the candidate information; The step of screening the plurality of information summary cards according to the target user portrait data and the plurality of card tags to obtain at least one content card to be recommended includes: According to the target user portrait data, a target virtual resource identifier of a target virtual resource is obtained, wherein the target virtual resource includes a virtual resource associated with the target user; According to the target virtual resource identifier, determining at least one initial content card from the plurality of information summary cards through the mapping relationship; The at least one to-be-recommended content card is determined from the at least one initial content card according to the card label of the at least one initial content card.

3. The method according to claim 2, characterized in that The determining, according to the card label of the at least one initial content card, the at least one to-be-recommended content card from the at least one initial content card comprises: Determine filtering rule data associated with the target user based on the target user portrait data, wherein the filtering rule data includes at least one of the following: historical recommended content, language preference information, region information, and information browsing time information; The at least one initial content card is filtered according to the filtering rule data and the card label of the at least one initial content card to determine the at least one content card to be recommended.

4. The method according to claim 2, characterized in that: Before acquiring the target user portrait data of the target user, the method further includes: Obtain the virtual resources that the user is interested in and stored in the data center to obtain the virtual resource list data; storing the virtual resource list data in a public database; Correspondingly, obtaining a target virtual resource identifier of a target virtual resource according to the target user portrait data includes: According to the user identifier in the target user portrait data, acquiring the target virtual resource that the target user is interested in from the virtual resource list data in the public database; A target virtual resource identifier of the target virtual resource is determined.

5. The method according to any one of claims 1 to 4, characterized in that: The screening of the plurality of financial information to obtain a plurality of candidate information includes: Determining financial information that meets a preset condition among the plurality of financial information as the plurality of candidate information; The preset conditions include at least one of the following: the number of words in the information belongs to the word number range, the information category belongs to the preset category, there are related virtual resources, there is a viewpoint tendency, the timeliness of the information belongs to the timeliness range, the review status is passed, and the information display status is available for external use. The display and information visible area includes the area of ​​the target user.

6. The method according to any one of claims 1 to 4, characterized in that: The step of pushing the at least one content card to be recommended to the application client of the target user includes: Determine the information recommendation feed flow and target insertion position; At the target insertion position of the information recommendation feed stream, the at least one content card to be recommended is added, and the added information recommendation feed stream is pushed to the application client.

7. The method according to any one of claims 1 to 4, characterized in that: Before pushing the at least one content card to be recommended to the application client of the target user, the method includes: Determine the display order of each of the at least one content card to be recommended based on display attribute information of the at least one content card to be recommended, wherein the display attribute information includes at least one of the following: information release time, information consultation word count, information like number, information forwarding number, and information title first letter; The step of pushing the at least one content card to be recommended to the application client of the target user includes: According to the display order of each of the at least one content card to be recommended, the at least one content card to be recommended is pushed to the application client.

8. The method according to any one of claims 1 to 4, characterized in that: Before pushing the at least one content card to be recommended to the application client of the target user, the method includes: Acquire a historical user corresponding to each of the at least one content card to be recommended, wherein the historical user refers to a user who has clicked or viewed the corresponding content card to be recommended; Clustering the user portrait data of at least one of the historical users, and obtaining a cluster center as target user feature data corresponding to each of the at least one content card to be recommended; Determining a recommendation degree for recommending the at least one to-be-recommended content card to the application client according to the target user portrait data of the target user and at least one of the target user feature data; Sorting the content cards to be recommended according to the recommendation degrees corresponding to the content cards to be recommended, and selecting the content cards to be recommended with a preset number of rankings at the top; The step of pushing the at least one content card to be recommended to the application client of the target user includes: The to-be-recommended content cards with a preset number of values ​​in the front ranking are used as information summary cards inserted into the community recommendation flow and pushed to the application client of the target user.

9. A content recommendation device, characterized in that: include, The first screening module is used to screen multiple financial information to obtain multiple candidate information; A first generating module is used to generate, for any candidate information among the plurality of candidate information, an information summary card of the candidate information and a card label of the information summary card according to the information content of the candidate information; A first acquisition module is used to acquire target user portrait data of a target user; A second screening module is used to screen the multiple information summary cards according to the target user portrait data and the multiple card tags to obtain at least one content card to be recommended; The push module is used to push the at least one content card to be recommended to the application client of the target user.

10. An electronic device, characterized in that: include: processor; as well as A memory, configured to store executable instructions of the processor; The processor is configured to perform the method of any one of claims 1 to 8 by executing the executable instructions.

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

Images