Content recommendation method and system based on digital financial AI platform

Abstract

The invention discloses a content recommendation method and system based on a digital financial AI platform, belongs to the technical field of artificial intelligence, and aims to solve the problem that an existing content recommendation method of a digital financial platform cannot meet the personalized and real-time requirements of a user. Comprising the following steps: acquiring historical behavior data of a first user on a digital financial AI platform; performing semantic analysis on the historical behavior data by using a natural language processing technology to obtain a semantic analysis result; constructing a first financial behavior semantic matrix of the first user according to the semantic analysis result; according to the first financial behavior semantic matrix of the first user, automatically retrieving and acquiring related financial news and analysis report data at preset time intervals; and sending the financial news and the analysis report data to terminal equipment of the first user.

Description

Technical Field

[0001] The present invention relates to the field of artificial intelligence technology, and specifically to a content recommendation method and system based on a digital finance AI platform. Background Art

[0002] In the current era of rapid development of digital information, the dynamics and information in the financial market have become increasingly complex and diverse. For investors, timely access to accurate and valuable financial news and analysis reports is the key to making reasonable investment decisions. However, in the face of a vast amount of information resources, users often need to spend a lot of time and effort to retrieve and filter by themselves to find the latest financial news related to the investment fields or companies they are interested in. This cumbersome process is not only inefficient but also may lead to missing important market dynamics, thus affecting the timeliness and accuracy of investment decisions.

[0003] Although existing digital finance platforms can provide rich financial news and analysis reports, there is still much room for improvement in content recommendation. Most of these platforms rely on users to manually search or push based on simple keyword matching, unable to meet the personalized and real-time needs of users.

[0004] The disclosure of the above background art content is only for assisting in understanding the concept and technical solution of the present invention, and it does not necessarily belong to the prior art of this patent application. Without clear evidence indicating that the above content was publicly available on the filing date of this patent application, the above background art should not be used to evaluate the novelty and inventiveness of this application. Summary of the Invention

[0005] This application provides a content recommendation method and system based on a digital finance AI platform to solve the problem that the content recommendation method of existing digital finance platforms cannot meet the personalized and real-time needs of users.

[0006] To achieve the above object, the embodiments of this application disclose the following technical solutions:

[0007] In a first aspect, the embodiments of this application provide a content recommendation method based on a digital finance AI platform, including the following steps:

[0008] Obtain the historical behavior data of a first user on the digital finance AI platform;

[0009] Use natural language processing technology to perform semantic analysis on the historical behavior data to obtain a semantic analysis result;

[0010] Construct a first financial behavior semantic matrix of the first user according to the semantic analysis result;

[0011] Retrieve and obtain relevant financial news and analysis report data at preset time intervals according to the first financial behavior semantic matrix of the first user;

[0012] Send the financial news and analysis report data to the terminal device of the first user.

[0013] In the embodiments of the present application, by using natural language processing (NLP) technology to perform semantic analysis on the historical behavior data of the user, the complex semantic relationships in the user's historical behavior data can be understood, so as to accurately extract the specific interests of the user, and further ensure that the subsequent recommended content has a high degree of relevance to the user's interests. By structuring the semantic analysis results into a financial behavior semantic matrix, it is convenient for quick access and calculation. In this way, on the one hand, the efficiency of data processing can be improved; on the other hand, it can also make the content matching and retrieval more accurate. By automatically retrieving the latest financial news and analysis reports at preset time intervals, the timeliness of the recommended content can be ensured. In this way, the problem that the content recommendation method of the existing digital financial platform cannot meet the personalized and real-time needs of users is effectively solved.

[0014] In some possible implementation manners of the first aspect, the historical behavior data includes information on the user's position holding, and the information on the user's position holding includes at least the asset name and the purchase price;

[0015] The steps of performing semantic analysis on the historical behavior data by using natural language processing technology to obtain the semantic analysis results include:

[0016] Perform word segmentation on the information on the user's position holding;

[0017] Perform named entity recognition on the segmented data, identify the entity category and generate the corresponding annotation;

[0018] Use predefined regular expression pattern matching to identify the asset name and purchase price recognized, and extract the corresponding text information and numerical information;

[0019] According to the asset name and purchase price of each holding item extracted, generate multiple first behavior vectors to obtain the semantic analysis results;

[0020] The steps of constructing the first financial behavior semantic matrix according to the semantic analysis results include:

[0021] Generate a first preliminary financial behavior semantic matrix according to the multiple first behavior vectors;

[0022] Sort the first behavior vectors of the financial behavior semantic matrix in descending order according to the purchase price to generate the first financial behavior semantic matrix.

[0023] Among them, word segmentation can ensure that data processing is more standardized. Identifying entity categories in text and labeling them can improve the understanding and processing capabilities of user position information. Next, using predefined regular expression patterns to match the identified asset names and purchase prices can efficiently and accurately extract the corresponding text information and numerical information.

[0024] After generating a preliminary financial behavior semantic matrix based on the extracted multiple first behavior vectors, the final first financial behavior semantic matrix is ​​generated by sorting the purchase prices from high to low, which can give priority to displaying users' large investments in certain assets, making it easier to allocate resources to information retrieval and recommendation of high-value holdings. In this way, on the one hand, system resources can be used more effectively and the overall recommendation efficiency can be improved; on the other hand, since users are often more interested in relevant information about high-value holdings, it is also conducive to improving user satisfaction; on the other hand, high-value holdings are often accompanied by higher risks. By highlighting these high-risk holdings and allocating resources to information retrieval and recommendation of high-value holdings, users can also make more robust investment decisions.

[0025] In some possible implementations of the first aspect, according to the first financial behavior semantic matrix, the step of automatically retrieving and acquiring relevant financial news and analysis report data at preset time intervals includes:

[0026] Get the list of first financial news websites and the list of first financial analysis report platforms;

[0027] Constructing a dynamic search request according to the asset name information in the first behavior vector in the first financial behavior semantic matrix in the sorted order;

[0028] Send the constructed dynamic search request to each financial news website in the first financial news website list and each financial analysis report platform in the first financial analysis report platform list, and crawl the returned web page content;

[0029] Based on the publishing timestamps contained in the web page content, the crawled web page content is time-filtered and the web page content within the preset time period is retained;

[0030] Clean and analyze the screened webpage content to extract financial news data and financial analysis report data;

[0031] Send the extracted financial news data and financial analysis report data to the terminal device of the first user. Among them, according to the asset name information in the user behavior matrix, a dynamic retrieval request is constructed, which can ensure that the crawled content is highly relevant. Through the screening of web page content crawled based on the publication timestamp, it can ensure that users obtain the latest financial news and analysis reports, thus further meeting the user's demand for real-time information.

[0032] In some possible implementation manners of the first aspect, the steps of obtaining the first financial news website list and the first financial analysis report platform list include:

[0033] Obtain a predefined financial news website list and a predefined financial analysis report platform list;

[0034] According to the asset name information in the first behavior vector, screen the predefined financial news website list and the predefined financial analysis report platform list;

[0035] Use the screened financial news websites and financial analysis report platforms as the target data sources;

[0036] Based on the access interface protocol of the target data source, establish an access connection to the target data source, verify the validity of the access connection, and retain the valid financial news websites and financial analysis report platforms to obtain the first financial news website list and the first financial analysis report platform list. In this way, on the one hand, the most relevant data sources can be screened out, which is conducive to improving the relevance and accuracy of the recommended content; on the other hand, it can avoid wasting resources on those inaccessible or low-quality data sources, thereby improving the efficiency and quality of data crawling.

[0037] In some possible implementation manners of the first aspect, the steps of obtaining the first financial news website list and the first financial analysis report platform list further include:

[0038] Receive the financial news websites and financial analysis report platforms manually input by the user;

[0039] Verify the data format and validity of the user input;

[0040] Add the verified user input data to the predefined financial news website list and the predefined financial analysis report platform list. In this way, users can manually input data sources, websites or platforms that they consider to be of high quality and high relevance. When making content recommendations based on these user-specified data sources, more accurate and personalized financial news and analysis reports can be provided, thereby effectively improving the relevance of the recommended content and user satisfaction.

[0041] In some possible embodiments of the first aspect, the historical behavior data further includes financial content information collected and / or shared by the first user; the content recommendation method based on the digital financial AI platform further includes the following steps:

[0042] Perform word segmentation on the financial content information collected and / or shared by the first user to generate a preliminary keyword list;

[0043] Count the occurrence frequency of each keyword in the preliminary keyword list, and select the keywords whose occurrence frequency exceeds a preset threshold to obtain a set of high-frequency keywords;

[0044] Generate multiple second behavior vectors according to each extracted keyword and the corresponding occurrence frequency of each keyword;

[0045] Generate a second preliminary financial behavior semantic matrix according to the multiple second behavior vectors;

[0046] Sort the second behavior vectors of the financial behavior semantic matrix in descending order of occurrence frequency to generate a second financial behavior semantic matrix of the first user;

[0047] Obtain a list of the first financial news websites and a list of the first financial analysis report platforms;

[0048] Construct a dynamic retrieval request in sequence according to the keywords in the second behavior vectors in the second financial behavior semantic matrix in the sorted order;

[0049] Send the constructed dynamic retrieval request to each financial news website in the list of the first financial news websites and each financial analysis report platform in the list of the first financial analysis report platforms to crawl the returned web content;

[0050] Based on the published timestamps included in the web content, perform time filtering on the crawled web content to retain the web content within a preset time period;

[0051] Perform data cleaning and parsing on the filtered web content, and extract the financial news data and financial analysis report data therein;

[0052] Send the extracted financial news data and financial analysis report data to the terminal device of the first user. In this way, the dimension of the user behavior data can be enriched, the scope of analysis can be expanded, which is conducive to more in-depth understanding of the user's interests and preferences, thereby greatly improving the accuracy of recommendation.

[0053] In a second aspect, an embodiment of the present application provides a content recommendation system based on a digital financial AI platform, including:

[0054] A first acquisition module, configured to acquire historical behavior data of a first user on the digital financial AI platform;

[0055] A semantic analysis module, which is used to perform semantic analysis on historical behavior data by using natural language processing technology to obtain a semantic analysis result;

[0056] A first construction module, which is used to construct a first financial behavior semantic matrix of the first user according to the semantic analysis result;

[0057] A first retrieval module, which is used to retrieve and obtain relevant financial news and analysis report data by itself at preset time intervals according to the first financial behavior semantic matrix of the first user;

[0058] A first content recommendation module, which is used to send the financial news and analysis report data to the terminal device of the first user.

[0059] In some possible implementation manners of the second aspect, the historical behavior data includes user's position holding information, and the user's position holding information includes at least the asset name and the purchase price;

[0060] Specifically, the semantic analysis module is used for: performing word segmentation processing on the user's position holding information; performing named entity recognition on the segmented data to identify entity categories and generate corresponding annotations; using a predefined regular expression pattern to match the recognized asset name and purchase price, and extracting the corresponding text information and numerical information; generating multiple first behavior vectors according to the asset name and purchase price of each holding item extracted, to obtain a semantic analysis result; the steps of constructing a first financial behavior semantic matrix according to the semantic analysis result include: generating a first preliminary financial behavior semantic matrix according to the multiple first behavior vectors; sorting each first behavior vector of the financial behavior semantic matrix in descending order of the purchase price to generate a first financial behavior semantic matrix.

[0061] In some possible implementation manners of the second aspect, the first retrieval module is specifically used for: obtaining a list of first financial news websites and a list of first financial analysis report platforms; constructing a dynamic retrieval request in sequence according to the asset name information in the first behavior vectors in the first financial behavior semantic matrix in the sorting order; sending the constructed dynamic retrieval request to each financial news website in the list of first financial news websites and each financial analysis report platform in the list of first financial analysis report platforms, and crawling the returned web page content; performing time screening on the crawled web page content based on the publication timestamps included in the web page content, and retaining the web page content within a preset time period; performing data cleaning and parsing on the screened web page content, and extracting the financial news data and financial analysis report data therein; sending the extracted financial news data and financial analysis report data to the terminal device of the first user.

[0062] In some possible embodiments of the second aspect, the first retrieval module is further specifically configured to: obtain a predefined list of financial news websites and a predefined list of financial analysis report platforms; screen the predefined list of financial news websites and the predefined list of financial analysis report platforms according to the asset name information in the first behavior vector; use the screened financial news websites and financial analysis report platforms as target data sources; establish an access connection to the target data sources based on the access interface protocols of the target data sources, verify the effectiveness of the access connection, and retain the valid financial news websites and financial analysis report platforms to obtain a first list of financial news websites and a first list of financial analysis report platforms.

[0063] In some possible embodiments of the second aspect, the first retrieval module is further specifically configured to: receive the financial news websites and financial analysis report platforms manually input by the user; verify the data format and effectiveness of the user input; add the verified user input data to the predefined list of financial news websites and the predefined list of financial analysis report platforms.

[0064] In some possible embodiments of the second aspect, the historical behavior data further includes financial content information collected and / or shared by the first user; the content recommendation system based on the digital finance AI platform further includes a second content recommendation module, which is configured to perform word segmentation on the financial content information collected and / or shared by the first user to generate a preliminary keyword list; count the occurrence frequency of each keyword in the preliminary keyword list, select the keywords with an occurrence frequency exceeding a preset threshold to obtain a set of high-frequency keywords; generate multiple second behavior vectors according to each extracted keyword and the corresponding occurrence frequency of each keyword; generate a second preliminary financial behavior semantic matrix according to the multiple second behavior vectors; sort the second behavior vectors of the financial behavior semantic matrix in descending order of occurrence frequency to generate a second financial behavior semantic matrix of the first user; obtain the first list of financial news websites and the first list of financial analysis report platforms; construct dynamic retrieval requests in sequence according to the keywords in the second behavior vectors in the second financial behavior semantic matrix in the sorted order; send the constructed dynamic retrieval requests to each financial news website in the first list of financial news websites and each financial analysis report platform in the first list of financial analysis report platforms to crawl the returned web page content; perform time screening on the crawled web page content based on the publication timestamps included in the web page content, and retain the web page content within a preset time period; perform data cleaning and parsing on the screened web page content, and extract the financial news data and financial analysis report data therein; send the extracted financial news data and financial analysis report data to the terminal device of the first user.

[0065] In a third aspect, an embodiment of the present application provides an electronic device, including one or more processors; a storage device storing one or more programs thereon; when the one or more programs are executed by the one or more processors, the one or more processors implement the method described in any technical solution of the first aspect.

[0066] In a fourth aspect, an embodiment of the present application provides a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, the method described in any technical solution of the first aspect is implemented.

[0067] In a fifth aspect, an embodiment of the present application provides a computer program product, including a computer program, and when the computer program is executed by a processor, the method described in any technical solution of the first aspect is implemented.

[0068] Among them, for the technical effects brought by any one of the design manners in the second aspect to the fifth aspect, reference may be made to the technical effects brought by different design manners in the first aspect, which will not be elaborated here. BRIEF DESCRIPTION OF THE DRAWINGS

[0069] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following will briefly introduce the drawings required for use in the description of the embodiments or the prior art. Obviously, the drawings in the following description are only exemplary, and for those of ordinary skill in the art, without creative efforts, other implementation drawings can also be obtained according to the provided drawings.

[0070] Figure 1 It is a schematic flowchart of a content recommendation method based on a digital finance AI platform provided by some embodiments of the present application;

[0071] Figure 2 It is a schematic structural diagram of a content recommendation system based on a digital finance AI platform provided by some embodiments of the present application;

[0072] Figure 3 It is a schematic structural diagram of an electronic device suitable for implementing some embodiments of the present application. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0073] Now, the specific implementation manners of the present invention will be described in detail. Although the present invention is described in conjunction with these specific implementation manners, it should be understood that it is not intended to limit the present invention to these specific implementation manners. On the contrary, these implementation manners are intended to cover alternative, changed, or equivalent implementation manners that may be included within the spirit and scope of the invention defined by the claims. In the following description, a large number of specific details are set forth in order to provide a comprehensive understanding of the present invention. The present invention can be implemented without some or all of these specific details.

[0074] As used in this specification and the appended claims, the singular forms "a", "an", and "the" include plural references unless the context clearly dictates otherwise. Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.

[0075] Overview of the Application: In the current era of rapid development of digital information, the dynamics and information in the financial market have become increasingly complex and diverse. For investors, timely access to accurate and valuable financial news and analysis reports is the key to making reasonable investment decisions. However, in the face of a vast amount of information resources, users often need to spend a large amount of time and effort to retrieve and filter on their own to find the latest financial news related to the investment fields or companies they are interested in. This cumbersome process is not only inefficient but may also lead to missing important market dynamics, thus affecting the timeliness and accuracy of investment decisions.

[0076] Although existing digital financial platforms can provide rich financial news and analysis reports, there is still much room for improvement in content recommendation. Most of these platforms rely on users to manually search or push based on simple keyword matching, unable to meet the personalized and real-time needs of users.

[0077] In response to the above technical problems, the general idea of the technical solution provided in this application is as follows: Provide a content recommendation method based on a digital financial AI platform, including the following steps: Obtain the historical behavior data of a first user on the digital financial AI platform; Use natural language processing technology to perform semantic analysis on the historical behavior data to obtain a semantic analysis result; Construct a first financial behavior semantic matrix of the first user according to the semantic analysis result; According to the first financial behavior semantic matrix of the first user, retrieve and obtain relevant financial news and analysis report data at preset time intervals; Send the financial news and analysis report data to the terminal device of the first user.

[0078] Among them, by using natural language processing (NLP) technology to perform semantic analysis on the historical behavior data of users, the complex semantic relationships in the historical behavior data of users can be understood, so as to accurately extract the specific interests of users, and then it can be ensured that the subsequent recommended content has a high degree of relevance to the interests of users. By structuring the semantic analysis results into a financial behavior semantic matrix, it is convenient for quick access and calculation. In this way, on the one hand, the efficiency of data processing can be improved; on the other hand, it can also make the content matching and retrieval more accurate. By automatically retrieving the latest financial news and analysis reports at preset time intervals, the timeliness of the recommended content can be ensured. In this way, the problem that the existing content recommendation methods of digital financial platforms cannot meet the personalized and real-time needs of users is effectively solved.

[0079] After introducing the basic principle of this application, the various non-limiting implementation manners of this application will be specifically introduced below in conjunction with the accompanying drawings of the specification.

[0080] Please refer to Figure 1 , the embodiment of this application provides a content recommendation method based on a digital financial AI platform, including the following steps:

[0081] S101: Obtain the historical behavior data of the first user on the digital financial AI platform.

[0082] Specifically, in some embodiments, the execution subject (for example, a computer device) of the content recommendation method based on the digital financial AI platform can be connected to the backend server of the digital financial AI platform through a wired connection or a wireless connection to obtain the historical behavior data of the first user on the digital financial AI platform. The historical behavior data may include, but is not limited to, user position information, and the user position information includes the asset name, purchase price, purchase date, and purchase quantity.

[0083] S102: Use natural language processing technology to perform semantic analysis on the historical behavior data to obtain a semantic analysis result.

[0084] Specifically, in some embodiments, the above execution subject can perform semantic analysis on the historical behavior data by using natural language processing technology through the following steps to obtain a semantic analysis result:

[0085] The first step is to perform word segmentation processing on the user position information;

[0086] Specifically, the above execution subject can perform word segmentation processing on the user position information through the Jieba word segmentation algorithm. Of course, this application is not limited to this. In other embodiments, the above execution subject can also perform word segmentation processing on the user position information through the Stanford word segmentation algorithm.

[0087] The second step is to perform named entity recognition on the segmented data, identify the entity category and generate corresponding annotations;

[0088] Specifically, the execution subject may use the BERT named entity recognition model to perform named entity recognition on the segmented data, and identify key entities from the segmented data, including but not limited to asset names and purchase prices. It is understandable that in other embodiments, pre-trained named entity recognition models such as spaCy and HanLP may also be used to identify key entities from the segmented data.

[0089] The third step is to use the predefined regular expression pattern to match the identified asset name and purchase price, and extract the corresponding text information and numerical information;

[0090] The regular expression and code examples are as follows:

[0091] import re

[0092] text="""

[0093] The user holds the following assets:

[0094] 1. Apple stock purchase price 145.30

[0095] 2. Tesla stock purchase price 720.50

[0096] 3. Google stock purchase price 2750.00

[0097] """

[0098] #Define regular expression pattern

[0099] asset_pattern = r"Stock\s([A-Za-z]+)\sPurchase price\s(\d+(.\d{1,2})?)"

[0100] #Use regular expressions to find matches

[0101] matches=re.findall(asset_pattern,text)

[0102] # The results are stored in a dictionary

[0103] portfolio=[]

[0104] for match in matches:

[0105] asset_name = match[0]

[0106] buy_price = float(match[1])

[0107] portfolio.append({"Asset Name": asset_name, "Buying Price": buy_price})

[0108] # Print the extraction result

[0109] for item in portfolio:

[0110] print(f"Asset Name: {item['Asset Name']}, Buying Price: {item['Buying Price']}")

[0111] Step 4: Based on the asset name and buying price of each held item extracted, generate multiple first-row vectors to obtain the semantic analysis result;

[0112] Specifically, the first-row vector includes the following elements: asset name text information and buying price numerical information.

[0113] Among them, performing word segmentation can ensure more standardized and normalized data processing. Identifying the entity categories in the text and performing annotation can improve the understanding and processing ability of the user's position information. Next, using the predefined regular expression pattern to match the identified asset name and buying price can efficiently and accurately extract the corresponding text information and numerical information.

[0114] S103: Construct the first financial behavior semantic matrix of the first user according to the semantic analysis result;

[0115] Specifically, the above execution entity can construct the first financial behavior semantic matrix of the first user through the following steps according to the semantic analysis result:

[0116] Step 1: Generate the first preliminary financial behavior semantic matrix according to multiple first-row vectors;

[0117] Among them, an example of the first preliminary financial behavior semantic matrix is as follows:

[0118] Asset Name Purchase Price (CNY) Company Name 1 150.00 Company Name 2 700.00 Company Name 3 3200.00

[0119] Step 2: Sort each first-row vector of the financial behavior semantic matrix in descending order according to the buying price to generate the first financial behavior semantic matrix.

[0120] After generating a preliminary financial behavior semantic matrix based on multiple extracted first behavior vectors, and then sorting it in descending order of the buying price to generate the final first financial behavior semantic matrix, it can preferentially display the large-scale investments of users in certain assets, facilitating subsequent information retrieval and recommendation for preferentially allocating resources to high-value positions. In this way, on the one hand, system resources can be utilized more effectively, improving the overall recommendation efficiency; on the other hand, since users tend to be more interested in information related to high-value positions, it is also conducive to improving user satisfaction; on the third hand, high-amount positions are often accompanied by higher risks. By highlighting these high-risk positions and preferentially allocating resources to information retrieval and recommendation for high-value positions, it can also help users make more prudent investment decisions.

[0121] S104: According to the first financial behavior semantic matrix of the first user, retrieve and obtain relevant financial news and analysis report data at preset time intervals.

[0122] Specifically, in some embodiments, the above-mentioned execution entity can retrieve and obtain relevant financial news and analysis report data at preset time intervals according to the first financial behavior semantic matrix through the following steps:

[0123] The first step is to obtain the first list of financial news websites and the first list of financial analysis report platforms.

[0124] Specifically, in some embodiments, the above-mentioned execution entity can obtain the first list of financial news websites and the first list of financial analysis report platforms through the following steps:

[0125] The first sub-step is to obtain the predefined list of financial news websites and the predefined list of financial analysis report platforms; specifically, these lists can include the URLs and access interface information of multiple well-known financial news websites and financial analysis report platforms.

[0126] The second sub-step is to screen the predefined list of financial news websites and the predefined list of financial analysis report platforms according to the asset name information in the first behavior vector.

[0127] Specifically, the above-mentioned execution entity can, according to the asset name information in each first behavior vector in the first financial behavior semantic matrix, match one by one the keywords marked by each financial news website in the predefined financial news website list and each financial analysis report platform in the financial analysis report platform list, and only retain the financial news websites and financial analysis report platforms that contain the asset name keywords to generate a list of candidate target data sources. Among them, the keywords marked by each financial news website in the predefined financial news website list and each financial analysis report platform in the financial analysis report platform list can be manually input and marked by staff, or can be obtained by accessing the corresponding website, crawling the content, and parsing and extracting keywords.

[0128] The third sub-step is to use the filtered financial news websites and financial analysis report platforms as the target data sources;

[0129] The fourth sub-step is to establish an access connection to the target data source based on the access interface protocol of the target data source, verify the validity of the access connection, and retain the valid financial news websites and financial analysis report platforms to obtain a list of the first financial news websites and a list of the first financial analysis report platforms. In this way, on the one hand, the most relevant data sources can be screened out, which is conducive to improving the relevance and accuracy of the recommended content; on the other hand, it can avoid wasting resources on those inaccessible or low-quality data sources, thereby improving the efficiency and quality of data crawling.

[0130] Preferably, on the basis of the above-mentioned embodiment, the above-mentioned execution entity can also obtain the list of the first financial news websites and the list of the first financial analysis report platforms through the following steps:

[0131] Receive the financial news websites and financial analysis report platforms manually input by the user;

[0132] Verify the data format and validity of the user input;

[0133] Add the user input data that passes the verification to the predefined financial news website list and the predefined financial analysis report platform list. In this way, users can manually input data sources, such as websites or platforms that they consider to be of high quality and high relevance. When making content recommendations based on these user-specified data sources, more accurate and personalized financial news and analysis reports can be provided, thereby effectively improving the relevance of the recommended content and user satisfaction. Of course, the present application is not limited to this. In other embodiments, it is also possible to directly access each website in the predefined financial news website list and the predefined financial analysis report platform list without screening the predefined financial news website list and the predefined financial analysis report platform list according to the asset name information in the first behavior vector, crawl the web content, and give priority to crawling the websites manually input by the user.

[0134] In the second step, in the order of sorting, construct a dynamic retrieval request according to the asset name information in the first row vector in the first financial behavior semantic matrix.

[0135] In the third step, send the constructed dynamic retrieval request to each financial news website in the list of first financial news websites and each financial analysis report platform in the list of first financial analysis report platforms, and crawl the returned web page content.

[0136] In the fourth step, based on the publication timestamps included in the web page content, perform time filtering on the crawled web page content, and retain the web page content within a preset time period; specifically, the preset time period is the absolute value of the time difference between the publication timestamp included in the web page content and the current retrieval time. For example, retain the web page content whose time difference between the publication timestamp and the current retrieval time is within 3 hours. It can be understood that the specific time depends on the requirements. In other embodiments, it can also be 8 hours, 12 hours, 24 hours, etc.

[0137] In the fifth step, perform data cleaning and parsing on the filtered web page content, and extract the financial news data and financial analysis report data therein.

[0138] Specifically, in some embodiments, the above execution subject can parse the cleaned web page content through the following steps using natural language processing technology to extract effective financial news data and financial analysis report data. The parsing process includes, but is not limited to, sub-steps such as information extraction, classification, and keyword extraction.

[0139] In the sixth step, send the extracted financial news data and financial analysis report data to the terminal device of the first user. Among them, constructing a dynamic retrieval request according to the asset name information in the user behavior matrix can ensure that the crawled content is highly relevant. By filtering the web page content crawled based on the publication timestamp, it can ensure that the user obtains the latest financial news and analysis reports, thereby further meeting the user's need for real-time information.

[0140] S105: Send the financial news and analysis report data to the terminal device of the first user.

[0141] Specifically, the push notification function of the mobile application can be used to push relevant content to the user's mobile phone or tablet in real time, or send a detailed news and report summary or full text link via email.

[0142] In some embodiments, the obtained historical behavior data further includes the financial content information collected and / or shared by the first user; the content recommendation method based on the digital finance AI platform further includes the following steps:

[0143] In the first step, perform word segmentation on the financial content information collected and / or shared by the first user to generate a preliminary keyword list;

[0144] In the second step, count the occurrence frequency of each keyword in the preliminary keyword list, and select the keywords whose occurrence frequency exceeds the preset threshold to obtain a set of high-frequency keywords;

[0145] In the third step, generate multiple second row vectors according to each extracted keyword and the corresponding occurrence frequency of each keyword;

[0146] In the fourth step, generate a second preliminary financial behavior semantic matrix based on the multiple second row vectors;

[0147] In the fifth step, sort the second row vectors of the financial behavior semantic matrix in descending order of occurrence frequency to generate the second financial behavior semantic matrix of the first user;

[0148] In the sixth step, obtain the list of the first financial news websites and the list of the first financial analysis report platforms;

[0149] Specifically, obtaining the list of the first financial news websites and the list of the first financial analysis report platforms can be designed in the above-described manner, which will not be elaborated here.

[0150] In the seventh step, construct a dynamic retrieval request in sequence according to the keywords in the second row vectors in the second financial behavior semantic matrix in the sorting order;

[0151] In the eighth step, send the constructed dynamic retrieval request to each financial news website in the list of the first financial news websites and each financial analysis report platform in the list of the first financial analysis report platforms, and crawl the returned web page content;

[0152] In the ninth step, based on the published timestamps included in the web page content, perform time screening on the crawled web page content, and retain the web page content within the preset time period;

[0153] In the tenth step, perform data cleaning and parsing on the screened web page content, and extract the financial news data and financial analysis report data therein;

[0154] In the eleventh step, send the extracted financial news data and financial analysis report data to the terminal device of the first user. In this way, the dimension of user behavior data can be enriched, the scope of analysis can be expanded, which is conducive to understanding the user's interests and preferences more deeply, thereby greatly improving the accuracy of recommendation.

[0155] Please refer to Figure 2, based on the same inventive concept as the content recommendation method based on a digital finance AI platform in the foregoing embodiments, an embodiment of the present application provides a content recommendation system based on a digital finance AI platform, including:

[0156] A first acquisition module 201, configured to acquire historical behavior data of a first user on the digital finance AI platform;

[0157] A semantic analysis module 202, configured to perform semantic analysis on the historical behavior data by using natural language processing technology to obtain a semantic analysis result;

[0158] A first construction module 203, configured to construct a first financial behavior semantic matrix of the first user according to the semantic analysis result;

[0159] A first retrieval module 204, configured to retrieve and acquire relevant financial news and analysis report data by itself at every preset time interval according to the first financial behavior semantic matrix of the first user;

[0160] A first content recommendation module 205, configured to send the financial news and analysis report data to the terminal device of the first user.

[0161] In some embodiments, the historical behavior data includes user's position holding situation information, and the user's position holding situation information at least includes the asset name and the purchase price;

[0162] The semantic analysis module 202 is specifically configured to: perform word segmentation processing on the user's position holding situation information; perform named entity recognition on the data after word segmentation, recognize the entity category and generate corresponding annotations; use a predefined regular expression pattern to match the recognized asset name and purchase price, and extract the corresponding text information and numerical information; generate a plurality of first behavior vectors according to the asset name and purchase price of each holding item extracted, to obtain a semantic analysis result; the steps of constructing a first financial behavior semantic matrix according to the semantic analysis result include: generating a first preliminary financial behavior semantic matrix according to the plurality of first behavior vectors; sorting each first behavior vector of the financial behavior semantic matrix in descending order of the purchase price to generate a first financial behavior semantic matrix.

[0163] In some embodiments, the first retrieval module 204 is specifically configured to: obtain a list of first financial news websites and a list of first financial analysis report platforms; sequentially construct dynamic retrieval requests according to the asset name information in the first row vectors in the first financial behavior semantic matrix in the sorting order; send the constructed dynamic retrieval requests to each financial news website in the list of first financial news websites and each financial analysis report platform in the list of first financial analysis report platforms, and crawl the returned web page content; based on the publication timestamps included in the web page content, perform time filtering on the crawled web page content, and retain the web page content within a preset time period; perform data cleaning and parsing on the filtered web page content, extract the financial news data and financial analysis report data therein; and send the extracted financial news data and financial analysis report data to the terminal device of the first user.

[0164] In some embodiments, the first retrieval module 204 is specifically further configured to: obtain a predefined list of financial news websites and a predefined list of financial analysis report platforms; filter the predefined list of financial news websites and the predefined list of financial analysis report platforms according to the asset name information in the first row vector; use the filtered financial news websites and financial analysis report platforms as target data sources; establish an access connection to the target data sources based on the access interface protocols of the target data sources, verify the validity of the access connection, and retain the valid financial news websites and financial analysis report platforms to obtain a list of first financial news websites and a list of first financial analysis report platforms.

[0165] In some embodiments, the first retrieval module 204 is specifically further configured to: receive the financial news websites and financial analysis report platforms manually input by the user; verify the data format and validity of the user input; and add the user input data that passes the verification to the predefined list of financial news websites and the predefined list of financial analysis report platforms.

[0166] In some embodiments, the historical behavior data further includes financial content information collected and / or shared by the first user; the content recommendation system based on the digital finance AI platform further includes a second content recommendation module 206, which is used to perform word segmentation on the financial content information collected and / or shared by the first user to generate a preliminary keyword list; count the occurrence frequency of each keyword in the preliminary keyword list, and select the keywords whose occurrence frequency exceeds a preset threshold to obtain a high-frequency keyword set; generate multiple second behavior vectors according to each extracted keyword and the corresponding occurrence frequency; generate a second preliminary financial behavior semantic matrix according to the multiple second behavior vectors; sort the second behavior vectors of the financial behavior semantic matrix in descending order of occurrence frequency to generate a second financial behavior semantic matrix of the first user; obtain a list of the first financial news websites and a list of the first financial analysis report platforms; construct a dynamic retrieval request in sequence according to the keywords in the second behavior vectors in the second financial behavior semantic matrix in the sorting order; send the constructed dynamic retrieval request to each financial news website in the list of the first financial news websites and each financial analysis report platform in the list of the first financial analysis report platforms, and crawl the returned web page content; based on the publication timestamp included in the web page content, perform time screening on the crawled web page content, and retain the web page content within a preset time period; perform data cleaning and parsing on the screened web page content, and extract the financial news data and financial analysis report data therein; send the extracted financial news data and financial analysis report data to the terminal device of the first user.

[0167] It can be understood that the various modules described in the content recommendation system based on the digital finance AI platform correspond to the respective steps in the content recommendation method based on the digital finance AI platform described in the reference Figure 1 Therefore, the operations, features, and beneficial effects described above for the method also apply to the content recommendation system based on the digital finance AI platform and the modules included therein, and will not be repeated here.

[0168] Please refer to Figure 3, based on the inventive concept of a content recommendation method based on a digital finance AI platform in the foregoing embodiments, an embodiment of the present application provides an electronic device. The electronic device may include, but is not limited to, mobile terminals such as mobile phones, laptop computers, digital broadcast receivers, PDAs (Personal Digital Assistants), PADs (Tablet Computers), PMPs (Portable Multimedia Players), etc., and fixed terminals such as digital TVs, desktop computers, etc. The electronic device includes a processing device 301 (such as a central processing unit, a graphics processing unit, etc.), which can perform various appropriate actions and processes according to a program stored in the ROM 302 (Read Only Memory) or a program loaded from the storage device 308 into the RAM 303 (Random Access Memory). In the RAM 303, various programs and data required for the operation of the electronic device are also stored. The processing device 301, the ROM 302, and the RAM 303 are connected to each other through a bus 304. An input / output interface (i.e., the I / O interface 305) is also connected to the bus 304.

[0169] Generally, the following devices can be connected to the I / O interface 305: an input device 306 including, for example, a touch screen, a touchpad, a keyboard, a mouse, a camera, a microphone, an accelerometer, a gyroscope, etc.; an output device 307 including, for example, a liquid crystal display (LCD), a speaker, a vibrator, etc.; a storage device 308 including, for example, a magnetic tape, a hard disk, etc.; and a communication device 309. The communication device 309 can allow the electronic device to communicate with other devices wirelessly or wiredly to exchange data.

[0170] Specifically, according to some embodiments of the present application, the process described above with reference to the flowchart can be implemented as a computer software program. For example, some embodiments of the present application include a computer program product, which includes a computer program carried on a computer-readable medium, and the computer program contains program codes for performing the method shown in the flowchart. In such some embodiments, the computer program can be downloaded and installed from the network through the communication device 309, or installed from the storage device 308, or installed from the ROM 302. When the computer program is executed by the processing device 301, the above functions defined in the method of some embodiments of the present application are executed.

[0171] It should be noted that the computer-readable medium described in some embodiments of the present application may be a computer-readable signal medium, a computer-readable storage medium, or any combination of the two. The computer-readable storage medium may be an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination of the above. More specific examples of the computer-readable storage medium may include, but are not limited to: an electrical connection having one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In some embodiments of the present application, the computer-readable storage medium may be any tangible medium that contains or stores a program, which can be used by or in conjunction with an instruction execution system, apparatus, or device. In some embodiments of the present application, the computer-readable signal medium may include a data signal propagated in a baseband or as part of a carrier wave, which carries computer-readable program code. Such a propagated data signal may take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. The computer-readable signal medium may also be any computer-readable medium other than the computer-readable storage medium, which can send, propagate, or transmit a program for use by or in conjunction with an instruction execution system, apparatus, or device. The program code contained on the computer-readable medium may be transmitted using any appropriate medium, including but not limited to: wires, optical cables, RF (radio frequency), etc., or any suitable combination of the above.

[0172] In some embodiments, the client and the server can communicate using any currently known or future-developed network protocol such as HTTP (HyperText Transfer Protocol), and can be interconnected with digital data communication in any form or medium (e.g., a communication network). Examples of communication networks include local area networks ("LAN"), wide area networks ("WAN"), the Internet (e.g., the Internet), and end-to-end networks (e.g., ad hoc end-to-end networks), as well as any currently known or future-developed network.

[0173] The above computer-readable medium may be included in the above electronic device; or it may exist separately without being assembled into the electronic device. The above computer-readable medium carries one or more programs. When the one or more programs are executed by the electronic device, the electronic device is caused to: obtain the historical behavior data of a first user on the digital finance AI platform; perform semantic analysis on the historical behavior data by using natural language processing technology to obtain a semantic analysis result; construct a first financial behavior semantic matrix of the first user according to the semantic analysis result; retrieve and obtain relevant financial news and analysis report data at preset time intervals according to the first financial behavior semantic matrix of the first user; and send the financial news and analysis report data to the terminal device of the first user.

[0174] Computer program code for performing the operations of some embodiments of the present application may be written in one or more programming languages or combinations thereof. The above programming languages include object-oriented programming languages such as Java, Smalltalk, C++, and also include conventional procedural programming languages such as the "C" language or similar programming languages. The program code may be executed entirely on the user's computer, partially on the user's computer, executed as a stand-alone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In the case of a remote computer, the remote computer may be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (for example, by using an Internet service provider to connect through the Internet).

[0175] The flowcharts and block diagrams in the accompanying drawings illustrate the possible architectures, functions, and operations of systems, methods, and computer program products according to various embodiments of the present application. In this regard, each block in the flowchart or block diagram may represent a module, a program segment, or a part of code that contains one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions marked in the blocks may occur in a different order than marked in the accompanying drawings. For example, two consecutive blocks shown may actually be executed substantially in parallel, and they may sometimes be executed in the reverse order, depending on the functions involved. It should also be noted that each block in the block diagram and / or flowchart, and the combinations of blocks in the block diagram and / or flowchart, may be implemented by a dedicated hardware-based system for performing the specified functions or operations, or may be implemented by a combination of dedicated hardware and computer instructions.

[0176] The modules described in some embodiments of the present application can be implemented in software or in hardware. The described modules can also be provided in a processor. For example, they can be described as: the first acquisition module 201, the semantic analysis module 202, the first construction module 203, the first retrieval module 204, the first content recommendation module 205, and the second content recommendation module 206.

[0177] Among them, the names of these modules do not constitute a limitation on the modules themselves in some cases. For example, the first acquisition module can also be described as the "historical behavior data acquisition module".

[0178] The functions described above in this article can be at least partially executed by one or more hardware logic components. For example, without limitation, the exemplary types of hardware logic components that can be used include: Field Programmable Gate Array (FPGA), Application Specific Integrated Circuit (ASIC), Application Specific Standard Product (ASSP), System on Chip (SOC), Complex Programmable Logic Device (CPLD), and so on.

[0179] Some embodiments of the present application also provide a computer program product, including a computer program, which when executed by a processor implements any one of the above content recommendation methods based on a digital finance AI platform.

[0180] Although the present invention has been described in detail above with general descriptions and specific embodiments, based on the present invention, some modifications or improvements can be made, which are obvious to those skilled in the art. Therefore, these modifications or improvements made without departing from the spirit of the present invention all fall within the scope of protection required by the present invention.

Claims

1. A content recommendation method based on a digital financial AI platform, characterized in that: The following steps are involved: Obtain the historical behavior data of the first user on the digital financial AI platform; Using natural language processing technology to perform semantic analysis on the historical behavior data to obtain semantic analysis results; Constructing a first financial behavior semantic matrix of the first user according to the semantic analysis result; According to the first financial behavior semantic matrix of the first user, automatically retrieve and obtain relevant financial news and analysis report data at preset time intervals; The financial news and the analysis report data are sent to a terminal device of the first user.

2. The content recommendation method based on the digital financial AI platform according to claim 1, characterized in that: The historical behavior data includes user position information, and the user position information at least includes the asset name and purchase price; The steps of performing semantic analysis on the historical behavior data using natural language processing technology to obtain semantic analysis results include: Performing word segmentation processing on the user's position information; Perform named entity recognition on the segmented data to identify the entity category and generate corresponding annotations; Use predefined regular expression patterns to match the identified asset names and purchase prices, and extract the corresponding text and numerical information; Generate multiple first behavior vectors according to the asset name and purchase price of each held item extracted, and obtain the semantic analysis result; The step of constructing the first financial behavior semantic matrix according to the semantic analysis result comprises: generating a first preliminary financial behavior semantic matrix according to a plurality of first behavior vectors; The first behavior vectors of the financial behavior semantic matrix are sorted according to the order of the purchase price from high to low to generate the first financial behavior semantic matrix.

3. The content recommendation method based on the digital financial AI platform according to claim 2, characterized in that: According to the first financial behavior semantic matrix, the steps of automatically retrieving and acquiring relevant financial news and analysis report data at preset time intervals include: Get the list of first financial news websites and the list of first financial analysis report platforms; Constructing a dynamic search request according to the asset name information in the first behavior vector in the first financial behavior semantic matrix in a sorted order; Sending the constructed dynamic search request to each financial news website in the first financial news website list and each financial analysis report platform in the first financial analysis report platform list, crawling the returned web page content; Based on the publishing timestamps contained in the web page content, the crawled web page content is time-filtered and the web page content within the preset time period is retained; Clean and analyze the screened webpage content to extract financial news data and financial analysis report data; The extracted financial news data and financial analysis report data are sent to the terminal device of the first user.

4. The content recommendation method based on the digital financial AI platform according to claim 3 is characterized in that: The steps to obtain the list of first financial news websites and the list of first financial analysis report platforms include: Get a list of predefined financial news websites and a list of predefined financial analysis report platforms; Filtering a predefined list of financial news websites and a predefined list of financial analysis report platforms according to the asset name information in the first behavior vector; Select the selected financial news websites and financial analysis report platforms as target data sources; Based on the access interface protocol of the target data source, an access connection to the target data source is established, the validity of the access connection is verified, and valid financial news websites and financial analysis report platforms are retained to obtain the first financial news website list and the first financial analysis report platform list.

5. The content recommendation method based on the digital financial AI platform according to claim 4, characterized in that: The steps to obtain the list of first financial news websites and the list of first financial analysis report platforms also include: Financial news websites and financial analysis report platforms that receive manual input from users; Verify the format and validity of data entered by users; Add the validated user input data to the predefined list of financial news websites and the predefined list of financial analysis report platforms.

6. The content recommendation method based on a digital financial AI platform according to any one of claims 1 to 5, characterized in that: The historical behavior data also includes financial content information collected and / or shared by the first user; the content recommendation method based on the digital financial AI platform also includes the following steps: Performing word segmentation processing on the financial content information collected and / or shared by the first user to generate a preliminary keyword list; Counting the frequency of occurrence of each keyword in the preliminary keyword list, selecting keywords whose frequency of occurrence exceeds a preset threshold, and obtaining a high-frequency keyword set; Generate multiple second behavior vectors according to each extracted keyword and the occurrence frequency corresponding to each keyword; generating a second preliminary financial behavior semantic matrix according to a plurality of second behavior vectors; Sorting the second behavior vectors of the financial behavior semantic matrix according to the order of occurrence frequency from high to low to generate a second financial behavior semantic matrix of the first user; Get the list of first financial news websites and the list of first financial analysis report platforms; constructing a dynamic search request according to the keywords in the second behavior vector in the second financial behavior semantic matrix in a sorted order; Sending the constructed dynamic search request to each financial news website in the first financial news website list and each financial analysis report platform in the first financial analysis report platform list, crawling the returned web page content; Based on the publishing timestamps contained in the web page content, the crawled web page content is time-filtered and the web page content within the preset time period is retained; Clean and analyze the screened webpage content to extract financial news data and financial analysis report data; The extracted financial news data and financial analysis report data are sent to the terminal device of the first user.

7. A content recommendation system based on a digital financial AI platform, characterized in that: include: A first acquisition module, used to acquire historical behavior data of the first user on the digital financial AI platform; A semantic analysis module, used to perform semantic analysis on the historical behavior data using natural language processing technology to obtain semantic analysis results; A first construction module, configured to construct a first financial behavior semantic matrix of a first user according to the semantic analysis result; A first retrieval module, configured to retrieve and obtain relevant financial news and analysis report data at preset time intervals according to the first financial behavior semantic matrix of the first user; The first content recommendation module is used to send the financial news and the analysis report data to the terminal device of the first user.

8. An electronic device, characterized in that: include: one or more processors; a storage device having one or more programs stored thereon; When the one or more programs are executed by the one or more processors, the one or more processors implement the method according to any one of claims 1 to 6.

9. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and when the computer program is executed by a processing device, the method according to any one of claims 1 to 6 is implemented.

10. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processing device, the method according to any one of claims 1 to 6 is implemented.

Images