Financial content recommendation method and device, electronic equipment and storage medium

Abstract

The application provides a financial content recommendation method and device, electronic equipment and storage medium, and relates to the field of financial technology or other related technical fields. The method comprises the following steps: in response to an information input operation triggered by a user on a recommendation prompt page, obtaining query information corresponding to the information input operation; obtaining at least one multimedia content matched with the query information and determining label information corresponding to each of the at least one multimedia content; determining target label information matched with personalized information of the user according to the label information corresponding to each of the at least one multimedia content; generating a financial recommendation page according to target multimedia content corresponding to the target label information, and displaying the financial recommendation page. The method improves the accuracy and effectiveness of financial content recommendation.

Description

Technical Field

[0001] This application relates to the field of financial technology or other related technical fields, and in particular to a method, apparatus, electronic device and storage medium for recommending financial content. Background Technology

[0002] With the rapid development of internet technology, the financial sector is shifting from offline counter services to online digital services. Customers in the financial sector are increasingly demanding access to and browsing of financial information.

[0003] Currently, a common query strategy in the financial field is to retrieve content matching the user's query text and then provide that content to the user for viewing. However, this method of directly recommending results based on text queries does not result in high click-through rates from users, leading to low accuracy in the recommendations.

[0004] Therefore, how to accurately provide users with recommended content is an urgent problem to be solved in the current financial field. Summary of the Invention

[0005] This application provides a financial content recommendation method, apparatus, electronic device, and storage medium to address the problem in the financial field where the click-through rate of recommended content is not high, resulting in low recommendation accuracy.

[0006] Firstly, this application provides a method for recommending financial content, including:

[0007] In response to a user's information input operation triggered by the recommendation prompt page, obtain query information corresponding to the information input operation;

[0008] Obtain at least one multimedia content that matches the query information and determine the tag information corresponding to at least one multimedia content;

[0009] Based on the tag information corresponding to at least one multimedia content, determine the target tag information that matches the user's personalized information;

[0010] Based on the target multimedia content corresponding to the target tag information, a financial recommendation page is generated and displayed.

[0011] Secondly, this application provides a financial content recommendation device, comprising:

[0012] The query acquisition unit is used to respond to the information input operation triggered by the user on the recommendation prompt page and obtain query information corresponding to the information input operation;

[0013] The content query unit is used to obtain at least one multimedia content that matches the query information and determine the tag information corresponding to each of the at least one multimedia content.

[0014] The content management unit is used to determine target tag information that matches the user's personalized information based on the tag information corresponding to at least one multimedia content.

[0015] The content recommendation unit is used to generate a financial recommendation page based on the target multimedia content corresponding to the target tag information, and to display the financial recommendation page.

[0016] Thirdly, this application provides an electronic device, including: a processor, and a memory communicatively connected to the processor;

[0017] The memory stores computer-executed instructions;

[0018] The processor executes computer execution instructions stored in the memory to implement the financial content recommendation method as described in the first aspect.

[0019] Fourthly, this application provides a computer-readable storage medium storing computer-executable instructions, which, when executed by a processor, are used to implement the financial content recommendation method as described in the first aspect.

[0020] Fifthly, this application provides a computer program product, including a computer program that, when executed by a processor, implements the financial content recommendation method described in the first aspect.

[0021] The solution provided in this application can respond to a user's information input operation triggered by a recommendation prompt page, obtain query information corresponding to the information input operation, and obtain at least one multimedia content matching the query information, and determine the tag information corresponding to each of the at least one multimedia content. Compared with the original text content query, multimedia content is richer. For the multimedia content obtained from the query, user personalized information can be used to obtain target tag information matching the personalized information from the multimedia content, making the target tag information match the user's personalized information. Furthermore, the financial recommendation page generated based on the target multimedia content corresponding to the target tag information is more closely matched to the user's personalized needs, thereby achieving targeted personalized recommendations. This makes the financial recommendation page displayed to the user more closely match the user's needs, achieving the goal of accurate content recommendation in the financial field. Attached Figure Description

[0022] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this application and, together with the description, serve to explain the principles of this application.

[0023] Figure 1 An application scenario diagram for implementing a financial content recommendation method is provided in the embodiments of this application;

[0024] Figure 2 A flowchart illustrating yet another embodiment of a financial content recommendation method provided in this application;

[0025] Figure 3 A flowchart illustrating yet another embodiment of a financial content recommendation method provided in this application;

[0026] Figure 4 A flowchart illustrating yet another embodiment of a financial content recommendation method provided in this application;

[0027] Figure 5 A flowchart illustrating yet another embodiment of a financial content recommendation method provided in this application;

[0028] Figure 6 A schematic diagram of one embodiment of a financial content recommendation device provided in this application;

[0029] Figure 7 This is a block diagram of an electronic device for implementing a financial content recommendation method, provided as an embodiment of this application.

[0030] The accompanying drawings illustrate specific embodiments of this application, which will be described in more detail below. These drawings and descriptions are not intended to limit the scope of the concept in any way, but rather to illustrate the concept of this application to those skilled in the art through reference to particular embodiments. Detailed Implementation

[0031] Exemplary embodiments will now be described in detail, examples of which are illustrated in the accompanying drawings. When the following description relates to the drawings, unless otherwise indicated, the same numbers in different drawings denote the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this application. Rather, they are merely examples of apparatuses and methods consistent with some aspects of this application as detailed in the appended claims.

[0032] It should be noted that the content recommendation method, apparatus, electronic device and storage medium provided in this application can be applied to the fintech field, or to any field other than fintech. This application does not limit the application field of the content recommendation method, apparatus, electronic device and storage medium.

[0033] The following is a detailed introduction to the existing technology:

[0034] Currently, content recommendation in the financial sector typically employs text-based queries. This method involves users entering query text on a bank's query page; this text can be keywords, short phrases, etc. The backend server retrieves the user's query text, searches for relevant content, and displays the retrieved content—for example, as a webpage or pop-up. However, this method generally relies on directly querying the user's input text; the probability of a user clicking and viewing the content is low. In other words, directly recommending content based on text queries may not match the user's actual needs, resulting in low accuracy.

[0035] To address the aforementioned issues and enhance the relevance of recommended content to user needs, thereby achieving more accurate content recommendations, the technical solution of this application considers, upon receiving a user's query request, obtaining query information corresponding to the input operation, and acquiring at least one multimedia content matching the query information, while determining the tag information corresponding to each multimedia content. Compared to traditional text-based queries, multimedia content offers greater richness. Furthermore, for the obtained multimedia content, user-specific information can be utilized to extract target tag information matching that personalization, ensuring a match between the target tag information and the user's personalized information. Consequently, the financial recommendation page generated based on the target multimedia content corresponding to the target tag information is more closely aligned with the user's personalized needs, achieving targeted personalized recommendations. This results in a financial recommendation page displayed to the user that better matches their requirements, improving the effectiveness and accuracy of the financial recommendation page.

[0036] The technical solution of this application and how the technical solution of this application solves the above-mentioned technical problems are described in detail below with specific embodiments. These specific embodiments can be combined with each other, and the same or similar concepts or processes may not be described again in some embodiments. The embodiments of this application will now be described with reference to the accompanying drawings.

[0037] like Figure 1 The diagram shown illustrates an application scenario for implementing a financial content recommendation method, as provided in an embodiment of this application. Figure 1 As shown, the scenario may include a user terminal 1, a backend server 2 that has a communication connection with the user terminal 1, and a content database 3 that is connected to the backend server.

[0038] In this embodiment, user terminal 1 can display a query page, which may include input controls. The system detects user input actions triggered by the input controls and sends the input actions to backend server 2. Backend server 2 can be a computer, cloud server, ordinary server, supercomputer, or other similar devices. In this embodiment, the specific device type of backend server 2 is not specifically limited.

[0039] The backend server 2 can receive information input operations sent by user terminal 1. In response to these input operations, it obtains the query information from the input operation. It then retrieves at least one multimedia content matching the query information and determines the tag information corresponding to each multimedia content, for example, such as... Figure 1 The multimedia content shown is 1-n. Each multimedia content can be associated with tag information. The multimedia content and its tag information can be stored in content database 3. Then, target tag information matching the user's personalized information can be determined from the multimedia content in content database 3. Furthermore, using the target multimedia content corresponding to the target tag information, a financial recommendation page can be generated. Since the target tag information matches the user's personalized information, the corresponding target multimedia content is more suitable for the user's needs. The financial recommendation page generated based on the target multimedia content can achieve more accurate content recommendations.

[0040] like Figure 2 The diagram shown is a flowchart of one embodiment of a financial content recommendation method provided in this application. The financial content recommendation method may include the following steps:

[0041] 201. In response to the user's information input operation triggered by the recommendation prompt page, obtain the query information corresponding to the information input operation.

[0042] Optionally, before step 201, the process may include receiving an information input operation, where the information input operation refers to an input operation for querying information triggered by a user.

[0043] Furthermore, the detection step of the information input operation may include: detecting the query information entered by the user, and generating an information input operation based on the query information. The information input operation can refer to an operation determined based on the user's query information.

[0044] The execution entity in this embodiment can be an electronic device, which can be a backend server or a user terminal; this embodiment does not limit this. For example, when the electronic device is a user terminal, it can directly detect and query content. When the electronic device is a server, it can receive information input operations sent by the user terminal.

[0045] In step 201, obtaining the query information corresponding to the information input operation may include parsing the information input operation to obtain the query information contained in the information input operation.

[0046] The query information can be information that represents the user's query needs. The query information can include any one or more of the following types of information: keywords, sentences, voice signals, images, videos, short videos, etc.

[0047] 202. Obtain at least one multimedia content that matches the query information and determine the tag information corresponding to each of the at least one multimedia content.

[0048] Optionally, obtaining at least one multimedia content that matches the query information may include querying at least one multimedia content associated with the query information through a content query algorithm.

[0049] Furthermore, multimedia content can be retrieved from a multimedia database. A multimedia database can refer to a database used to store multimedia content. Specifically, at least one piece of multimedia content matching the query information can be retrieved from the multimedia database.

[0050] This involves identifying keywords for multimedia content in a multimedia database, matching the query information with the keywords of each multimedia content, and obtaining at least one multimedia content that matches the query information.

[0051] In this embodiment, obtaining at least one multimedia content matching the query information may include obtaining at least one multimedia content matching the query information through an acquisition model. The acquisition model may include generative artificial intelligence (AIGC).

[0052] Multimedia content can include one or more of the following types of content: text, images, audio, and video.

[0053] 203. Based on the tag information corresponding to at least one multimedia content, determine the target tag information that matches the user's personalized information.

[0054] Optionally, the tag information for multimedia content refers to the content tags set for the multimedia content. These tags can be used to simply record the content type of the multimedia content. The tag information for multimedia content can include one or more tag keywords. For example, the tag information for multimedia information could be the words "animals" and "scenery," while the tag information for another movie video could be the words "romance" and "suspense."

[0055] Step 203 may include: calculating the relevance between the tag information and the user's personalized information based on the tag information corresponding to at least one multimedia content, obtaining the relevance of the tag information, and determining the target tag information that meets the relevant conditions based on the relevance of the tag information.

[0056] Furthermore, meeting the relevant conditions can mean that the relevance is greater than or equal to a relevance threshold. Relevance can refer to the degree of correlation between tag information and user personalized information, and specifically, the relevance can be a value greater than or equal to 0 and less than or equal to 1.

[0057] This process of calculating the relevance between tag information and user personalized information can include: inputting the tag information and user personalized information into a machine learning model for relevance calculation; using the machine learning model to predict the relevance between the tag information and the user personalized information; and obtaining the relevance between the tag information and the user personalized information. Specifically, the machine learning model can be a trained neural network model that can calculate the relevance between two input pieces of information and display the relevance between them.

[0058] In this embodiment, the user's personalized information can refer to information obtained by feature extraction or keyword extraction from data such as the user's interests, preferences, and behaviors. Personalized information may include personalized features or personalized keywords. Features, for example, can refer to word vectors (embeddings) corresponding to keywords. In this embodiment, personalized information is used to represent the user's aforementioned interests, preferences, and behaviors.

[0059] 204. Generate and display a financial recommendation page based on the target multimedia content corresponding to the target tag information.

[0060] Optionally, step 204 may include: generating a financial recommendation page based on the target multimedia content corresponding to the target tag information, according to the financial information display method. The financial information display method may include: the interaction type corresponding to the display channel of the financial information; for details regarding this display method, please refer to the descriptions in other embodiments. Furthermore, the financial information display method may also include: a preset information display method, such as a preset webpage format or pop-up format.

[0061] Taking the pop-up format as an example, the target multimedia content corresponding to the target tag information can be set in the video display area of ​​the pop-up window according to the video display area and title display area in the pop-up window format. Based on the content information corresponding to the target multimedia content, a content title can be generated and set in the title display area to obtain the financial recommendation pop-up window obtained at the end of the video and title setting, and then the financial recommendation pop-up window can be displayed.

[0062] The above display method of the financial recommendation page is merely an example and should not constitute a limitation on the specific display method of the financial recommendation page.

[0063] The financial content recommendation method provided in this application can respond to a user's information input operation triggered by a recommendation prompt page, obtain query information corresponding to the information input operation, and obtain at least one multimedia content matching the query information, and determine the tag information corresponding to each of the at least one multimedia content. Compared with the original text content query, multimedia content is richer. For the multimedia content obtained by the query, the user's personalized information can be used to obtain target tag information matching the personalized information from the multimedia content, so that the target tag information matches the user's personalized information. Then, the financial recommendation page generated according to the target multimedia content corresponding to the target tag information is more matched with the user's personalized needs, thereby realizing targeted personalized recommendations, making the financial recommendation page displayed to the user more matched with the user's needs, and achieving the goal of accurate content recommendation in the financial field.

[0064] As an optional implementation, obtaining at least one multimedia content matching the query information and determining the tag information corresponding to each of the at least one multimedia content includes:

[0065] Load at least one initial acquisition model obtained through training. Each initial acquisition model is used to acquire different types of multimedia content, and includes at least one of a text acquisition model, an image acquisition model, and a video acquisition model.

[0066] Based on pre-set financial characteristic information, at least one initial acquisition model is updated to obtain the target acquisition model.

[0067] Input the query information into at least one target acquisition model and obtain the acquisition content output by at least one target acquisition model.

[0068] Optionally, loading at least one initial acquisition model obtained through training may include: calling an acquisition model that has been trained to obtain at least one initial acquisition model.

[0069] Optionally, at least one initial acquisition model includes at least one of a text acquisition model, an image acquisition model, and a video acquisition model.

[0070] The pre-set financial characteristic information may include at least one of the following: financial-related field type information, content playback type, application scenario type, etc.

[0071] In this embodiment, updating at least one initial acquisition model based on pre-set financial characteristic information to obtain a target acquisition model may include: updating the parameters of at least one initial acquisition model based on pre-set financial characteristic information and query information to obtain a target acquisition model, making the target acquisition model more suitable for financial characteristic information and query information, thereby improving the quality and efficiency of content generation.

[0072] Furthermore, the model parameters of at least one initial acquisition model can be fine-tuned and adapted based on preset financial characteristic information and query information, so that the acquisition model has a higher adaptability to financial characteristic information and query information.

[0073] Taking parameter 'a' in the model as an example, parameter 'a' can be adjusted to parameter 'a', which is more relevant to the application scenario of financial characteristic information and query information, so that the target acquisition model corresponding to parameter 'a' has a higher degree of fit with financial characteristic information and query information.

[0074] Taking the display format of multimedia content as an example, if the query information is "How is the stock market today?", it can be determined that the application scenario corresponding to the query information is a real-time query scenario of financial content. The parameter a of the display format of multimedia content: the default text display format can be changed to parameter a': video display format, which is more compatible with the "real-time query scenario of financial content".

[0075] The financial content recommendation method provided in this application first obtains at least one initial acquisition model through training. Then, based on pre-set financial characteristic information, the initial acquisition model is updated, making the updated target acquisition model more suitable for content generation in the financial field, thus improving the quality and efficiency of the generated multimedia content. Furthermore, query information is input into at least one target acquisition model, yielding the acquisition content output by each model. This enables the querying of multimedia content more adapted to the characteristics of the financial field. The multimedia content obtained in this way is more applicable to the financial technology field, and using this multimedia content as the basis for recommendation further improves the effectiveness and accuracy of content recommendation.

[0076] like Figure 3 The flowchart shown is for another embodiment of a financial content recommendation method provided in this application. The difference from the previous embodiments is that, after obtaining at least one multimedia content matching the query information and determining the tag information corresponding to each of the at least one multimedia content, the method further includes:

[0077] 301. Based on the file format of the multimedia content, classify the multimedia content to obtain the category identifier of the multimedia content.

[0078] 302. Use content analysis algorithms to extract key information from multimedia content.

[0079] 303. Determine the attribute information of multimedia content based on its content source and time information.

[0080] 304. Generate tag information for multimedia content based on category identifiers, key information, and attribute information.

[0081] Optionally, the file format of multimedia content can refer to the storage format of multimedia content. For example, common image formats used to store images include bmp, jpg, png, tif, and gif; common file formats used to store text include word and txt; common video formats used to store video include AVI, WMV, and MPEG; and common audio signal formats include MIDI and WAV.

[0082] In this embodiment, the multimedia content is classified according to its file format to obtain a category identifier. This may include: determining the category to which the multimedia content belongs based on its file format; and generating a category identifier for the multimedia content based on its category.

[0083] The category identifier can be a unique identifier used to identify multimedia content, and it may include an identifier for the category to which the multimedia content belongs. Generating the category identifier for the multimedia content based on its category may include: determining the category number of the multimedia content's category and the content number corresponding to that category, and concatenating the category number and content number to obtain the multimedia content's category identifier.

[0084] In this embodiment, the key information in the multimedia content is extracted using a content analysis algorithm. This may include: using a content analysis model to extract information such as keywords, themes, and sentiment information from the multimedia content, and using the extracted keywords, themes, and sentiment information as the key information of the multimedia content.

[0085] Furthermore, content analysis algorithms may include one or more algorithms from technologies such as natural language processing, computer vision, and speech signal processing.

[0086] Optionally, determining the attribute information of the multimedia content based on its content source and time information may include: generating a source attribute for the multimedia content based on its content source, generating a time attribute for the multimedia content based on its time information, and determining the attribute information of the multimedia content based on the source attribute and time attribute. Further, the source attribute and time attribute can be determined as the attribute information of the multimedia content. In addition, other attributes of the multimedia content can be added to the attribute information; this embodiment does not impose excessive limitations on the specific content and quantity of the attribute information.

[0087] In this embodiment, the generation of tag information for multimedia content is based on the category identifier, key information, and attribute information of the multimedia content. This includes combining the category identifier, key information, and attribute information of the multimedia content to obtain the tag information of the multimedia content.

[0088] The financial content recommendation method provided in this application can classify multimedia content according to its file format to obtain category identifiers. Setting category tags facilitates the categorization and management of multimedia content. Furthermore, it can extract key information from the multimedia content based on content analysis algorithms for effective analysis. In addition, it can determine the attribute information of the multimedia content based on its source and time information; the attribute information represents the source and time of the multimedia content. Furthermore, by utilizing the category identifiers, key information, and attribute information of the multimedia content, tags can be set from multiple perspectives, enabling accurate identification of the multiple tags within the multimedia content. By effectively marking multimedia content with tags, and then utilizing these tags, user queries for multimedia content can be achieved, further improving the efficiency and accuracy of multimedia content queries.

[0089] As another optional implementation, based on the category identifier, key information, and attribute information of the multimedia content, tag information for the multimedia content is generated, including:

[0090] Obtain the tag structure used to record tag information;

[0091] The category identifiers, key information, and attribute information of multimedia content are converted into tag information corresponding to at least one multimedia content according to the tag structure.

[0092] A tag structure can refer to a data structure used to record tag information; that is, it can be a collection of data elements with specific relationships.

[0093] A tag structure can be, for example, a class with multiple data members, which can include, for example, category identifier members, key information members, attribute members, etc. When the information content is high, members can also include sub-members.

[0094] Optionally, converting the category identifier, key information, and attribute information of multimedia content into tag information corresponding to at least one multimedia content according to the tag structure may include: determining the category identifier member, key information member, and attribute member in the tag structure, setting the category identifier of the multimedia content in the category identifier member, setting the key information in the key information member, and setting the attribute information in the attribute member, so as to determine the tag information composed of the member information of each member.

[0095] The financial content recommendation method provided in this application can determine a tag structure for recording tag information and convert the category identifier, key information, and attribute information of multimedia content into tag information according to the tag structure. By setting the tag structure, the representation of tag information for different multimedia content can be made consistent. This allows for a unified filtering method when using the tag information of multimedia content to further improve the accuracy of the target multimedia content.

[0096] Another possible design also includes:

[0097] Store the multimedia content and the tag information corresponding to at least one multimedia content in the content database;

[0098] Based on the tag information corresponding to at least one multimedia content, determine the target tag information that matches the user's personalized information, including:

[0099] Based on the tag information of multimedia content stored in the content database, retrieve target tag information from the content database that matches the user's personalized information.

[0100] Optionally, after obtaining the multimedia content and its tag information, the multimedia content and its tag information can be associated and stored in a content database. The content database can include multiple multimedia content items, each of which can be associated with tag information.

[0101] The multimedia content stored in the content database can originate from the query results of multiple users' respective queries. By storing the multimedia content queried by multiple users in the content database, the number of multimedia contents participating in personalized information matching can be increased, the query scope can be expanded, and a larger amount of target multimedia content can be obtained, thus enriching the content of the financial recommendation page.

[0102] Based on the content database, target tag information that matches the user's personalized information is determined according to the tag information corresponding to at least one multimedia content. This may include: determining target tag information that matches the user's personalized information based on the tag information of multimedia content in the content database.

[0103] Furthermore, the similarity between the tag information of multimedia content in the content database and the user's personalized information can be calculated, and the tag information with the highest information similarity can be used as the target tag information to match the user's personalized information.

[0104] Optionally, the information similarity between tag information and personalized information can be obtained through feature extraction and feature distance calculation. Further, tag features can be extracted from the tag information, personalized features can be extracted from the personalized information, and the feature distance between the tag features and personalized features can be calculated to obtain the information similarity corresponding to the feature distance. Feature distance and information similarity are inversely proportional; the smaller the feature distance, the higher the information similarity. The larger the feature distance, the lower the information similarity.

[0105] The financial content recommendation method provided in this application stores multimedia content and at least one corresponding tag information in a content database. This content database allows for the associated storage of multimedia content and its tag information. When a query is needed, the content database is used to retrieve multimedia content and obtain more relevant target tag information. By setting up the content database, the amount of multimedia content available can be enriched, expanding the number of queries and resulting in more comprehensive and effective queries, ultimately leading to more accurate target tag information.

[0106] As another optional implementation, target tag information matching the user's personalized information is determined based on tag information corresponding to at least one multimedia content, including:

[0107] Based on a preset information collection strategy, personalized information of users is collected from multiple information processing systems;

[0108] Based on the user's personalized information and the tag information of multimedia content, calculate the relevance matrix between tag information and personalized information. The elements in the relevance matrix refer to the relevance between the tag information and personalized information corresponding to the position of the element.

[0109] Based on the relevance of each element in the relevance matrix, target elements whose relevance meets the preset relevance conditions are identified, and the label information corresponding to the target elements is determined as the target label information.

[0110] It is understood that before using the technical solutions disclosed in the various embodiments of this application, users should be informed of the types, scope of use, and usage scenarios of personal information and personalized information involved in this application in an appropriate manner in accordance with relevant laws and regulations, and user authorization should be obtained.

[0111] For example, upon receiving a user's proactive request, a prompt message is sent to the user to explicitly inform them that the requested operation will require the acquisition and use of the user's personalized information. This allows the user to autonomously choose, based on the prompt message, whether to provide personal information to the software or hardware such as the electronic device, application, server, or storage medium performing the operation of this application's technical solution.

[0112] As an optional but non-limiting implementation, in response to a user's active request, a prompt message can be sent to the user, for example, via a pop-up window, where the prompt message can be presented in text format. Furthermore, the pop-up window can also include a selection control for the user to choose "agree" or "disagree" to provide personalized information to the electronic device.

[0113] It is understood that the above notification and user authorization process is merely exemplary and does not limit the implementation of this application. Other methods that comply with relevant laws and regulations may also be applied to the implementation of this application.

[0114] Optionally, the information collection strategy may include: user profiling, collaborative filtering, association rules, and other technical strategies. Based on the preset information collection strategy, collecting personalized user information from multiple information processing systems may include: collecting user data from multiple information processing systems respectively, and using user profiling, collaborative filtering, association rules, and other technical strategies to determine the user's personalized information based on the user data corresponding to each of the multiple information processing systems.

[0115] In this embodiment, the multiple information processing systems can be multiple systems such as web browsing systems, information query systems, social systems, video browsing systems, shopping systems, and online education systems.

[0116] Optionally, multimedia content can include multiple elements; therefore, the tags corresponding to multimedia content can include multiple elements.

[0117] Based on the user's personalized information and the tag information of multimedia content, a relevance matrix between tag information and personalized information is calculated. This may include: reading multiple multimedia content items and their corresponding tag information from a content database; calculating the relevance between each tag information and the user's personalized information based on the tag information corresponding to each multimedia content item, obtaining the relevance between each tag information and the personalized information; and determining the relevance matrix based on the relevance between each tag information and the personalized information. Elements in the relevance matrix may include the relevance between the tag information corresponding to a specific meta-position and the user's personalized information.

[0118] Further, in this embodiment, determining the target elements whose relevance satisfies the preset relevance conditions may include: sorting the relevances corresponding to the elements in the relevance matrix, and selecting the top K elements as target elements; or, based on the relevances corresponding to the elements in the relevance matrix, selecting elements with relevances greater than a relevance threshold as target elements. If the number of selected elements with relevances greater than the relevance threshold is greater than K, then the top K elements with relevances greater than the relevance threshold and higher relevance values ​​can be selected as target elements. Here, K is a positive integer greater than 1.

[0119] The financial content recommendation method provided in this application can collect personalized information from multiple information processing systems according to a preset information collection strategy, expanding the scope of personalized information collection and obtaining more comprehensive and diverse personalized information. Utilizing the user's personalized information and the tag information of multimedia content, a relevance matrix is ​​calculated between the tag information and the personalized information. Elements in the relevance matrix represent the relevance between tags and personalized information at corresponding positions. A relative degree matrix is ​​used to calculate the relevance between tag information and personalized information, obtaining an accurate relevance representation. Then, from the relevance matrix, target elements whose relevance meets preset relevance conditions are selected, and the tag information corresponding to the target element is obtained as the target tag information. By using relevance as an evaluation criterion, the tag information selection process can be combined with the user's personalized information, resulting in a higher relevance between the obtained target tag information and the user's personalized information. By using the multimedia content corresponding to the target tag information as the target multimedia content, the target multimedia content can be more closely aligned with the user's personalized information, achieving more personalized content recommendations and a precise and effective recommendation effect.

[0120] Furthermore, based on any of the above embodiments, a relevance matrix between tag information and personalized information is calculated according to the user's personalized information and the tag information of multimedia content, including:

[0121] Identify at least one sub-information within a user's personalized information;

[0122] Calculate the relevance between the tag information of the multimedia content and each sub-information to obtain the sub-relevance between the tag information and at least one sub-information respectively;

[0123] The relevance between tag information and personalized information is determined based on the sub-relevance of tag information and at least one sub-information respectively.

[0124] Based on the correlation between tag information and personalized information, a correlation matrix is ​​determined for the corresponding tag information and personalized information.

[0125] Optionally, when calculating the relevance between tag information and personalized information, at least one sub-information in the personalized information can be identified, and the relevance between the tag information and each sub-information can be calculated based on the tag information of the multimedia content, so as to obtain the sub-relevance between the tag information and at least one sub-information respectively.

[0126] Furthermore, when calculating the sub-relevance between label information and each sub-information, at least one label member can be identified in the label information. The similarity of each member's corresponding information to the sub-information is then calculated to obtain the similarity between each label member and the sub-information. Finally, the weighted sum of these similarities yields the sub-relevance between the label information and the sub-information. By subdividing the label information, a more granular calculation of the relevance between the label information and each sub-information can be performed, resulting in more accurate calculation results.

[0127] In addition, when calculating the sub-relevance between label information and each sub-information, the label information can be converted into label features, and the sub-information features of each sub-information can be extracted. The feature similarity between the label features and the sub-information features can be calculated to obtain the sub-relevance between the label information and each sub-information.

[0128] In this embodiment, determining the relevance between tag information and personalized information based on the sub-relevance values ​​corresponding to tag information and at least one sub-information can include: weighted summing of the sub-relevance values ​​corresponding to tag information and at least one sub-information to obtain the relevance between tag information and personalized information. The weights corresponding to at least one sub-information can be set according to usage requirements. For example, if the information content corresponding to at least one sub-information is of equal importance, the same weight can be assigned to at least one sub-information. If the information content corresponding to at least one sub-information is of unequal importance, different weights can be assigned to at least one sub-information. The specific settings can be configured according to usage requirements.

[0129] The financial content recommendation method provided in this application, when calculating the relevance matrix of tag information and personalized information, can calculate the relevance of at least one sub-information in the personalized information with the tag information, respectively, to obtain the sub-relevance of the tag information with each of the at least one sub-information. The sub-relevance of the tag information with each of the at least one sub-information can determine the relevance between the tag information and the personalized information. Each sub-information can represent the user's personalized information from different angles or directions. By calculating the relevance of each sub-information with the tag information of multimedia content, an accurate relevance evaluation of the user's same tag information can be performed to obtain the relevance between the tag information and the personalized information. Furthermore, the relevance matrix is ​​determined based on the relevance between the tag information and the personalized information. By using different sub-information to represent personalized information, a multi-dimensional representation of personalized information can be achieved, increasing the evaluation dimensions of relevance and obtaining a more accurate relevance matrix.

[0130] like Figure 4 The flowchart shown is for another embodiment of a financial content recommendation method provided in this application. The difference from the previous embodiments is that, after displaying the financial recommendation page, it further includes:

[0131] 401. Collect the results of user actions performed on the financial recommendation page.

[0132] Optionally, the results of user actions on the financial recommendation page may include any one of the following: clicks, browsing or non-browsing results, browsing time, comments, reviews, etc.

[0133] Displaying a financial recommendation page can include showing the page to a user's terminal device. A user can also refer to a user categorized into a certain type or possessing certain special attributes. The financial recommendation page can also be sent to the terminal devices of other users belonging to the same category as the user, so that these other users' terminal devices can display the financial recommendation page.

[0134] 402. Based on the operation results, analyze and process the financial recommendation page to obtain the corresponding valuable content information.

[0135] Optionally, analyzing and processing the financial recommendation page based on the operation results to obtain the corresponding value content information may include: scoring the recommendation effect of the target multimedia content on the financial recommendation page based on the operation results to obtain the recommendation score of the target multimedia content, and using the recommendation score of the target multimedia content to determine the corresponding value content information of the financial recommendation page.

[0136] For example, the recommended scores of target multimedia content can be used to rank the content, and the ranked target multimedia content can be used as valuable content information.

[0137] For example, the recommendation scores of target multimedia content can be used for filtering to obtain target multimedia content with scores greater than a score threshold. The classification results of target multimedia content with scores greater than the score threshold and the remaining target multimedia content can be used as value content information.

[0138] 403. Update the acquisition model based on the value content information to obtain the updated acquisition model.

[0139] Optionally, step 403 may include: generating new training data based on the value content information, updating the acquisition model using the new training data, and obtaining the updated acquisition model. The acquisition model may include a target acquisition model, which uses the value content information to provide model feedback, enabling the target acquisition model to learn the user's actual operation results and improve the content acquisition accuracy of the target acquisition model.

[0140] 404. Based on the value content information, modify the multimedia content in the content database to obtain the modified content database.

[0141] Optionally, step 404 may include: performing operations such as adding, modifying, deleting, or merging multimedia content in the content database based on the value content information to obtain a modified content database.

[0142] Furthermore, based on the high-scoring target multimedia content in the value content information, operations such as adding, modifying, and merging multimedia content in the content database can be performed. Similarly, operations such as modifying and merging multimedia content in the content database can be performed based on the low-scoring target multimedia content in the value content information. High-scoring target multimedia content refers to target multimedia content with a score greater than a score threshold. Low-scoring target multimedia content refers to target multimedia content with a score less than or equal to a score threshold.

[0143] For example, adding multimedia content to the content database can mean querying new multimedia content similar to the high-scoring target multimedia content and adding the new multimedia content to the content database.

[0144] The financial content recommendation method provided in this application can obtain the user's operation results on the financial recommendation page. These operation results represent the user's feedback to the financial recommendation page. The method can then analyze and process these results to obtain the corresponding valuable content information. In other words, it uses the user's feedback to assess the recommendation effect of the financial recommendation page. Subsequently, this valuable content information is applied to updating the acquisition model and modifying the multimedia content in the content database. By influencing the acquisition model and content database with the user's feedback, the method makes the acquisition model and content database more closely aligned with the user's operation results. This improves the query performance and efficiency of the acquisition model and enhances the freshness of the content database, ensuring that the multimedia content in the content database maintains high quality.

[0145] As another alternative implementation, a financial recommendation page is displayed, including:

[0146] Obtain at least one recommendation channel from the financial recommendation page, and associate the recommendation channel with the corresponding application;

[0147] Based on the channel attribute information corresponding to at least one recommendation channel, determine the target recommendation channel that matches the user's usage habits.

[0148] Based on the applications associated with the target recommendation channels, the financial recommendation pages are converted into target recommendation content that corresponds to the interaction type of the applications;

[0149] Push the target recommended content to the application associated with the target recommendation channel so that the application associated with the target recommendation channel can display the target recommended content.

[0150] Optionally, at least one recommendation channel may include at least one of the following: SMS channels, email channels, social media channels, website channels, and application channels. Using one or more recommendation channels can improve the effectiveness of the financial recommendation page's display, avoid displaying the financial recommendation page on channels frequently used by users, reduce the probability of users viewing the financial recommendation page, and thus improve the recommendation effectiveness of the financial recommendation page.

[0151] Optionally, at least one recommendation channel may be associated with an application. Obtaining at least one recommendation channel for the financial recommendation page may include: obtaining the applications available on the user's terminal device, categorizing at least one application according to its application type, and obtaining at least one application corresponding to each type.

[0152] Channel attribute information can refer to attributes such as the frequency of use of the recommended channel, the source of the channel, and the duration of use.

[0153] Specifically, determining the target recommendation channels that match the user's usage habits based on the channel attribute information corresponding to at least one recommendation channel may include: calculating the probability of the user using the recommendation channels based on the channel attribute information corresponding to at least one recommendation channel, obtaining the usage probability corresponding to at least one recommendation channel, and determining the top N recommendation channels with the highest usage probability as the target recommendation channels that match the user's usage habits based on the usage probability corresponding to at least one recommendation channel.

[0154] Optionally, based on the application associated with the target recommendation, the financial recommendation page is converted into target recommendation content corresponding to the interaction type of the application, including: based on the application associated with the target recommendation, obtaining a template corresponding to the interaction type of the application, and converting the financial recommendation page into target recommendation content according to the template.

[0155] Optionally, pushing the target recommendation content to the application associated with the target recommendation channel includes: sending the target recommendation content to the application interface associated with the target recommendation channel based on the application interface of the application associated with the target recommendation channel, so as to push the target recommendation content to the application associated with the target recommendation channel.

[0156] The financial content recommendation method provided in this application can obtain at least one recommendation channel capable of displaying financial recommendation pages, with each recommendation channel associated with a corresponding application. By utilizing the channel attribute information corresponding to each of the at least one recommendation channel, a target recommendation channel matching the user's usage habits can be determined. Furthermore, when displaying target recommended content through the application associated with the target recommendation channel, the target recommended content is displayed through a more efficient channel, thereby improving the effectiveness of the target recommended content recommendation.

[0157] like Figure 5 The diagram shown is a flowchart illustrating a financial content recommendation method provided in an embodiment of this application. Figure 5 As shown, the financial content recommendation method provided in this embodiment is a more complete embodiment based on the financial content recommendation method provided in any of the above embodiments. The financial content recommendation method provided in this embodiment includes the following steps:

[0158] 501. In response to the user's information input operation triggered by the recommendation prompt page, obtain the query information corresponding to the information input operation.

[0159] 502. Obtain at least one multimedia content that matches the query information and determine the tag information corresponding to each of the at least one multimedia content.

[0160] 503. Based on the file format of the multimedia content, classify the multimedia content to obtain the category identifier of the multimedia content.

[0161] 504. Use content analysis algorithms to extract key information from multimedia content.

[0162] 505. Determine the attribute information of multimedia content based on its content source and time information.

[0163] 506. Generate tag information for multimedia content based on category identifiers, key information, and attribute information.

[0164] 507. Store the multimedia content and the tag information corresponding to at least one multimedia content in the content database.

[0165] 508. Based on the tag information of multimedia content stored in the content database, query the target tag information that matches the user's personalized information from the content database.

[0166] 509. Generate and display a financial recommendation page based on the target multimedia content corresponding to the target tag information.

[0167] The financial content recommendation method provided in this application can obtain query information corresponding to the user's information input operation after the user initiates the input operation, and then obtain multimedia content matching the query information. The multimedia content can be classified according to its file format to obtain category identifiers. Different multimedia content can be effectively distinguished using these category identifiers. Content analysis algorithms are also used to extract key information from the multimedia content, and this key information is used to provide a simple description of the multimedia content. Furthermore, the attribute information of the multimedia content can be determined based on its source and time information, enabling source tracking. Then, tag information for the multimedia content is generated using category identifiers, key information, and attribute information, achieving accurate labeling of the multimedia content. For the multimedia content obtained through the query, user personalized information can be used to obtain target tag information matching the personalized information from the multimedia content in the content database, ensuring a match between the target tag information and the user's personalized information. The content database can expand the query scope of multimedia content, obtaining a larger number of target tag information entries, and thus richer target multimedia content. The financial recommendation page generated based on the target multimedia content corresponding to the target tag information is more in line with the user's personalized needs and has a richer content, thereby achieving targeted, personalized, and comprehensive recommendations, and realizing accurate and effective content recommendations in the financial field.

[0168] like Figure 6The diagram shown is a structural schematic of one embodiment of a financial content recommendation device provided in this application. The financial content recommendation device 600 may include:

[0169] The query retrieval unit 601 is used to respond to the information input operation triggered by the user on the recommendation prompt page and obtain the query information corresponding to the information input operation.

[0170] Content query unit 602 is used to obtain at least one multimedia content that matches the query information and determine the tag information corresponding to each of the at least one multimedia content.

[0171] Content management unit 603 is used to determine target tag information that matches the user's personalized information based on the tag information corresponding to at least one multimedia content.

[0172] Content recommendation unit 604 is used to generate and display a financial recommendation page based on the target multimedia content corresponding to the target tag information.

[0173] As one embodiment, the content query unit includes:

[0174] The model loading module is used to load at least one initial acquisition model obtained through training; the at least one initial acquisition model is used to acquire different types of multimedia content, and the at least one initial acquisition model includes at least one of a text acquisition model, an image acquisition model, and a video acquisition model.

[0175] The model update module is used to update at least one initial acquisition model based on pre-set financial characteristic information to obtain the target acquisition model.

[0176] The conditional query module is used to input query information into at least one target acquisition model and obtain the acquisition content output by at least one target acquisition model respectively;

[0177] The content determination module is used to determine the acquired content output by at least one target acquisition model as at least one multimedia content.

[0178] As yet another embodiment, it also includes:

[0179] The identification setting unit is used to classify multimedia content according to its file format and obtain the category identification of the multimedia content.

[0180] The key extraction unit is used to extract key information from multimedia content using content analysis algorithms;

[0181] The attribute determination unit is used to determine the attribute information of multimedia content based on the content source and time information of the multimedia content.

[0182] The tag generation unit is used to generate tag information for multimedia content based on the category identifier, key information, and attribute information of the multimedia content.

[0183] As another embodiment, the label generation unit includes:

[0184] The structure acquisition module is used to acquire the tag structure used to record tag information.

[0185] The information conversion module is used to convert the category identifier, key information and attribute information of multimedia content into at least one tag information corresponding to each multimedia content according to the tag structure.

[0186] As yet another embodiment, it also includes:

[0187] An information storage unit is used to store multimedia content and tag information corresponding to at least one multimedia content into a content database;

[0188] The content management unit includes:

[0189] The content query module is used to retrieve target tag information that matches the user's personalized information from the content database, based on the tag information of multimedia content stored in the content database.

[0190] As another embodiment, the content management unit includes:

[0191] The information collection module is used to collect personalized user information from multiple information processing systems based on a preset information collection strategy.

[0192] The relevant calculation module is used to calculate the relevance matrix between tag information and personalized information based on the user's personalized information and the tag information of multimedia content. The elements in the relevance matrix refer to the relevance between the tag information and personalized information corresponding to the position of the element.

[0193] The element selection module is used to determine the target element whose relevance meets the preset relevance conditions based on the relevance of each element in the relevance matrix, and to determine the label information corresponding to the target element as the target label information.

[0194] As yet another embodiment, the relevant computing module is specifically used for:

[0195] Identify at least one sub-information in the user's personalized information; calculate the relevance between the tag information of the multimedia content and each sub-information to obtain the sub-relevance of the tag information and at least one sub-information respectively; determine the relevance between the tag information and the personalized information based on the sub-relevance of the tag information and at least one sub-information respectively; determine the relevance matrix between the tag information and the personalized information according to the relevance between the tag information and the personalized information.

[0196] As yet another embodiment, it also includes:

[0197] The operation collection unit is used to collect the results of user operations on the financial recommendation page;

[0198] The operation analysis unit is used to analyze and process the financial recommendation page based on the operation results to obtain the corresponding valuable content information of the financial recommendation page;

[0199] The model feedback unit is used to update the acquisition model based on the value content information to obtain the updated acquisition model;

[0200] The content feedback unit is used to modify the multimedia content in the content database based on valuable content information, and obtain the modified content database.

[0201] As another embodiment, the content recommendation unit includes:

[0202] The channel acquisition module is used to acquire at least one recommendation channel from the financial recommendation page, and the recommendation channel is associated with the corresponding application.

[0203] The channel determination module is used to determine the target recommendation channel that matches the user's usage habits based on the channel attribute information corresponding to at least one recommendation channel.

[0204] The channel conversion module is used to convert the financial recommendation page into target recommendation content corresponding to the interaction type of the application based on the application associated with the target recommendation channel.

[0205] The content recommendation module is used to push targeted recommended content to applications associated with the target recommendation channel, so that the applications associated with the target recommendation channel can display the targeted recommended content.

[0206] The financial content recommendation device provided in this embodiment can execute the technical solution of the financial content recommendation method provided in any of the above embodiments. Its implementation principle and technical effect are similar to those of the method embodiment shown in Embodiment 1, and will not be described in detail here.

[0207] Figure 7 This is a schematic diagram of the structure of an electronic device used to implement the financial content recommendation method in the embodiments of this application, such as... Figure 7 As shown, the electronic device 70 provided in this embodiment includes: a processor 72, and a memory 71 and a transceiver 73 that are communicatively connected to the processor 72;

[0208] Memory 71 stores computer-executed instructions; transceiver 73 is used for sending and receiving data;

[0209] The processor 72 executes computer execution instructions stored in the memory to implement the financial content recommendation method provided in any embodiment.

[0210] The memory 71, processor 72 and transceiver 73 are connected via bus 74.

[0211] The relevant explanations can be understood by referring to the relevant descriptions and effects corresponding to the steps of the financial content recommendation method provided in any embodiment, and will not be elaborated further here.

[0212] This invention also provides a computer-readable storage medium storing computer-executable instructions, which, when executed by a processor, are used to implement the financial content recommendation method provided in any embodiment.

[0213] This invention also provides a computer program product, including a computer program that is executed by a processor using the financial content recommendation method provided in any of the above embodiments.

[0214] It should be noted that, for the sake of simplicity, the foregoing method embodiments are all described as a series of actions. However, those skilled in the art should understand that this application is not limited to the described order of actions, as some steps may be performed in other orders or simultaneously according to this application. Furthermore, those skilled in the art should also understand that the embodiments described in the specification are all optional embodiments, and the actions and modules involved are not necessarily essential to this application.

[0215] It should be further noted that although the steps in the flowchart are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowchart may include multiple sub-steps or multiple stages. These sub-steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these sub-steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the sub-steps or stages of other steps.

[0216] It should be understood that the above-described device embodiments are merely illustrative, and the device of this application can also be implemented in other ways. For example, the division of units / modules in the above embodiments is only a logical functional division, and there may be other division methods in actual implementation. For example, multiple units, modules, or components may be combined, or integrated into another system, or some features may be ignored or not executed.

[0217] Furthermore, unless otherwise specified, the functional units / modules in the various embodiments of this application can be integrated into one unit / module, or each unit / module can exist physically separately, or two or more units / modules can be integrated together. The integrated units / modules described above can be implemented in hardware or as software program modules.

[0218] When integrated units / modules are implemented in hardware, the hardware can be digital circuits, analog circuits, etc. The physical implementation of the hardware structure includes, but is not limited to, transistors, memristors, etc. Unless otherwise specified, the processor can be any suitable hardware processor, such as a CPU, GPU, FPGA, DSP, and ASIC, etc. Unless otherwise specified, the storage unit can be any suitable magnetic or magneto-optical storage medium, such as Resistive Random Access Memory (RRAM), Dynamic Random Access Memory (DRAM), Static Random Access Memory (SRAM), Enhanced Dynamic Random Access Memory (EDRAM), High-Bandwidth Memory (HBM), Hybrid Memory Cube (HMC), etc.

[0219] If the integrated unit / module is implemented as a software program module and sold or used as an independent product, it can be stored in a computer-readable storage device (CMD). Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a memory and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of the various embodiments of this application. The aforementioned memory includes various media capable of storing program code, such as a USB flash drive, read-only memory (ROM), random access memory (RAM), portable hard drive, magnetic disk, or optical disk.

[0220] In the above embodiments, the descriptions of each embodiment have their own emphasis. For parts not described in detail in a certain embodiment, please refer to the relevant descriptions of other embodiments. The technical features of the above embodiments can be combined arbitrarily. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as the combination of these technical features does not contradict each other, it should be considered within the scope of this specification.

[0221] Other embodiments of this application will readily occur to those skilled in the art upon consideration of the specification and practice of the invention disclosed herein. This application is intended to cover any variations, uses, or adaptations of this application that follow the general principles of this application and include common knowledge or customary techniques in the art not disclosed herein. The specification and examples are to be considered exemplary only, and the true scope and spirit of this application are indicated by the following claims.

[0222] It should be understood that this application is not limited to the precise structure described above and shown in the accompanying drawings, and various modifications and changes can be made without departing from its scope. The scope of this application is limited only by the appended claims.

Claims

1. A method for recommending financial content, characterized in that, include: In response to a user's information input operation triggered by the recommendation prompt page, obtain query information corresponding to the information input operation; Obtain at least one multimedia content that matches the query information and determine the tag information corresponding to at least one multimedia content; Based on the tag information corresponding to at least one multimedia content, determine the target tag information that matches the user's personalized information; Based on the target multimedia content corresponding to the target tag information, a financial recommendation page is generated and displayed; The step of obtaining at least one multimedia content that matches the query information and determining the tag information corresponding to each of the at least one multimedia content includes: Load at least one initial acquisition model obtained through training, wherein the at least one initial acquisition model is used to acquire different types of multimedia content, and the at least one initial acquisition model includes at least one of a text acquisition model, an image acquisition model, and a video acquisition model; Based on pre-set financial characteristic information and query information, the parameters of the at least one initial acquisition model are updated to obtain the target acquisition model; The query information is input into at least one of the target acquisition models respectively to obtain the acquisition content output by at least one of the target acquisition models respectively; The content output by at least one of the target acquisition models is determined as at least one of the multimedia contents.

2. The method according to claim 1, characterized in that, After obtaining at least one multimedia content that matches the query information and determining the tag information corresponding to each of the at least one multimedia content, the method further includes: Based on the file format of the multimedia content, the multimedia content is classified to obtain the category identifier of the multimedia content; Key information in the multimedia content is extracted using content analysis algorithms; Based on the content source and time information of the multimedia content, determine the attribute information of the multimedia content; Based on the category identifier, key information, and attribute information of the multimedia content, tag information of the multimedia content is generated.

3. The method according to claim 2, characterized in that, The step of generating tag information for the multimedia content based on the category identifier, key information, and attribute information of the multimedia content includes: Obtain the tag structure used to record tag information; The category identifier, key information, and attribute information of the multimedia content are converted into tag information corresponding to at least one multimedia content according to the tag structure.

4. The method according to claim 2, characterized in that, Also includes: Store the multimedia content and the tag information corresponding to at least one multimedia content into the content database; The step of determining target tag information that matches the user's personalized information based on tag information corresponding to at least one multimedia content includes: Based on the tag information of multimedia content stored in the content database, target tag information that matches the user's personalized information is retrieved from the content database.

5. The method according to claim 1, characterized in that, The step of determining target tag information that matches the user's personalized information based on tag information corresponding to at least one multimedia content includes: Based on a preset information collection strategy, the user's personalized information is collected from multiple information processing systems; Based on the user's personalized information and the multimedia content's tag information, a correlation matrix between the tag information and the personalized information is calculated. The elements in the correlation matrix refer to the correlation between the tag information and the personalized information corresponding to the position of the element. Based on the relevance of each element in the relevance matrix, target elements whose relevance meets preset relevance conditions are determined, and the label information corresponding to the target elements is determined as the target label information.

6. The method according to claim 5, characterized in that, Based on the user's personalized information and the multimedia content's tag information, a relevance matrix between the tag information and the personalized information is calculated, including: Determine at least one sub-information in the user's personalized information; Calculate the relevance between the tag information of the multimedia content and each sub-information to obtain the sub-relevance between the tag information and at least one of the sub-information respectively; Based on the sub-relevance degrees corresponding to the tag information and at least one of the sub-information respectively, the relevance between the tag information and the personalized information is determined; Based on the relevance between the tag information and the personalized information, a relevance matrix corresponding to the tag information and the personalized information is determined.

7. The method according to claim 1, characterized in that, After displaying the financial recommendation page, the following is also included: Collect the results of the user's actions on the financial recommendation page; Based on the operation results, the financial recommendation page is analyzed and processed to obtain the value content information corresponding to the financial recommendation page; The acquisition model is updated based on the value content information to obtain the updated acquisition model; Based on the value content information, the multimedia content in the content database is modified to obtain the modified content database.

8. The method according to claim 1, characterized in that, The financial recommendation page includes: Obtain at least one recommendation channel from the financial recommendation page, wherein the recommendation channel is associated with a corresponding application; Based on the channel attribute information corresponding to each of the at least one recommendation channel, a target recommendation channel that matches the user's usage habit information is determined. Based on the application associated with the target recommendation channel, the financial recommendation page is converted into target recommendation content corresponding to the interaction type of the application; The target recommended content is pushed to the application associated with the target recommendation channel, so that the application associated with the target recommendation channel displays the target recommended content.

9. A financial content recommendation device, characterized in that, include: The query acquisition unit is used to respond to the information input operation triggered by the user on the recommendation prompt page and obtain query information corresponding to the information input operation; The content query unit is used to obtain at least one multimedia content that matches the query information and determine the tag information corresponding to each of the at least one multimedia content. The content management unit is used to determine target tag information that matches the user's personalized information based on the tag information corresponding to at least one of the multimedia contents; The content recommendation unit is used to generate a financial recommendation page based on the target multimedia content corresponding to the target tag information, and to display the financial recommendation page. The content query unit includes: The model loading module is used to load at least one initial acquisition model obtained through training. The at least one initial acquisition model is used to acquire different types of multimedia content. The at least one initial acquisition model includes at least one of a text acquisition model, an image acquisition model, and a video acquisition model. The model update module is used to update the parameters of the at least one initial acquisition model according to the pre-set financial characteristic information and query information to obtain the target acquisition model; The conditional query module is used to input the query information into at least one of the target acquisition models respectively, and obtain the acquisition content output by at least one of the target acquisition models respectively; The content determination module is used to determine the acquired content output by at least one of the target acquisition models as at least one of the multimedia contents.

10. An electronic device, comprising: A processor, and a memory communicatively connected to the processor; The memory stores computer-executed instructions; The processor executes computer execution instructions stored in the memory to implement the financial content recommendation method as described in any one of claims 1-8.

11. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer-executable instructions, which, when executed by a processor, are used to implement the financial content recommendation method as described in any one of claims 1 to 8.

Images