Analysis report generation method and related device

Abstract

The invention discloses an analysis report generation method and a related device. The method comprises the following steps: acquiring a template text, target basic data of a target bank operation mechanism and target operation data of the target bank operation mechanism; and based on the target basic data and the target operation data, processing through a report text generation model to obtain a first report text of the target bank operation institution. And generating a report chart of the target bank operation mechanism according to the first report text. And generating an analysis report of the target bank operating mechanism according to the template text, the first report text and the report chart. Therefore, semantic understanding is carried out on the target basic data and the target operation data of the report text generation model, the incidence relation and the context relation between the data are captured, and the target basic data and the target operation data are converted into the first report text capable of reflecting the characteristics of the target bank operation mechanism and the specific operation condition; therefore, a more targeted analysis report is generated in combination with the report chart and the template text.

Description

Technical Field

[0001] The present invention relates to the financial field, and in particular to a method for generating an analysis report and related devices. Background Art

[0002] Bank operating institutions usually summarize their operating conditions by writing analysis reports, and based on the analysis reports, guide future operating strategies and work plans.

[0003] In related technologies, an automated analysis tool is used to generate an analysis report according to a fixed template to reduce labor costs. However, the analysis report generated in this way is fixed and single, and it is difficult to generate a targeted analysis report by combining the characteristics of the bank operating institution and the actual operating conditions. Summary of the Invention

[0004] In view of the above problems, the present application provides a method for generating an analysis report and related devices, which are used to improve the pertinence and accuracy of the analysis report of bank operating institutions.

[0005] Based on this, the present application discloses the following technical solutions:

[0006] In a first aspect, an embodiment of the present application provides a method for generating an analysis report, the method including:

[0007] Obtain a template text, target basic data of a target bank operating institution, and target operating data of the target bank operating institution;

[0008] Based on the target basic data and the target operating data, perform processing through a report text generation model to obtain a first report text of the target bank operating institution;

[0009] Generate a report chart of the target bank operating institution according to the first report text;

[0010] Generate an analysis report of the target bank operating institution according to the template text, the first report text, and the report chart.

[0011] Optionally, the report text generation model includes basic data and historical operating data, and the performing processing through the report text generation model based on the target basic data and the target operating data to obtain the first report text of the target bank operating institution includes:

[0012] Determine the institution category of the target bank operating institution according to the target basic data and the basic data;

[0013] According to the institution category of the target bank operating institution, determine the same type of operating data of the bank operating institutions belonging to the institution category from the historical operating data;

[0014] Based on the differences between the target business data and the peer business data, process through the report text generation model to obtain the first report text of the target bank operating institution.

[0015] Optionally, the report text generation model is trained in the following manner:

[0016] Obtain target text sample pairs, where the target text sample pairs include bank operating institution samples and the target report texts corresponding to the bank operating institution samples;

[0017] Based on the bank operating institution samples, perform prediction through the initial report text generation model to obtain a predicted report text, where the initial report text generation model is used to predict the report text of the bank operating institution;

[0018] According to the differences between the predicted report text and the target report text, adjust the model parameters of the initial report text generation model to obtain the report text generation model.

[0019] Optionally, generating the report chart of the target bank operating institution based on the first report text includes:

[0020] Based on the first report text, perform identification through the chart generation model to obtain the target type of the report chart;

[0021] According to the target type, determine the target generation method of the report chart;

[0022] Through the chart generation model according to the target generation method, generate the report chart of the target bank operating institution.

[0023] Optionally, the method further includes:

[0024] Obtain external data, where the external data includes external economic data and map data;

[0025] Perform semantic recognition on the external data to obtain semantic information;

[0026] Based on the target basic data, the target business data, and the semantic information, process through the report text generation model to obtain the second report text of the target bank operating institution;

[0027] Generate the analysis report of the target bank operating institution according to the template text, the first report text, the second report text, and the report chart.

[0028] Optionally, the analysis report includes multiple chapters. After generating the analysis report of the target bank operating institution, the method further includes:

[0029] In response to obtaining a subscription request from a first user, determining a target chapter corresponding to the subscription request;

[0030] Splitting the analysis report according to the target chapter to obtain subscription content corresponding to the target chapter;

[0031] Sending the subscription content to the first user.

[0032] Optionally, after generating the analysis report of the target bank operating institution, the method further includes:

[0033] Generating a permission identifier for the analysis report, where the permission identifier is used to represent users who have the qualification to obtain the analysis report;

[0034] In response to obtaining a distribution request from a second user, matching the users represented by the permission identifier with the second user;

[0035] If the second user belongs to the users represented by the permission identifier, distributing the analysis report to the second user.

[0036] In a second aspect, an embodiment of the present application provides an analysis report generation device, and the device includes: an acquisition unit, a processing unit, and a generation unit;

[0037] The acquisition unit is configured to acquire a template text, target basic data of the target bank operating institution, and target operating data of the target bank operating institution;

[0038] The processing unit is configured to process the target basic data and the target operating data through a report text generation model to obtain a first report text of the target bank operating institution;

[0039] The generation unit is configured to generate a report chart of the target bank operating institution according to the first report text;

[0040] The generation unit is further configured to generate an analysis report of the target bank operating institution according to the template text, the first report text, and the report chart.

[0041] In a third aspect, an embodiment of the present application provides a computer device, and the computer device includes a processor and a memory:

[0042] The memory is configured to store a computer program and transmit the computer program to the processor;

[0043] The processor is configured to execute the method described in the above first aspect according to the computer program.

[0044] In a fourth aspect, an embodiment of the present application provides a computer-readable storage medium for storing a computer program, where the computer program is configured to execute the method described in the above first aspect.

[0045] In a fifth aspect, an embodiment of the present application provides a computer program product including a computer program, which, when running on a computer device, causes the computer device to execute the method described in the above first aspect.

[0046] From the above technical solutions, it can be seen that the present application has at least the following beneficial effects:

[0047] Obtain the template text, the target basic data of the target bank operating institution, and the target operating data of the target bank operating institution. Based on the target basic data and the target operating data, process them through the report text generation model to obtain the first report text of the target bank operating institution. Generate the report charts of the target bank operating institution according to the first report text. Generate the analysis report of the target bank operating institution according to the template text, the first report text, and the report charts. Thus, through semantic understanding of the target basic data that can reflect the basic situation of the target bank operating institution itself and the target operating data that can reflect the specific operating situation of the target bank operating institution by the report text generation model, capture the correlation and context relationships between the data, and convert the target basic data and the target operating data into the first report text that can reflect the characteristics and specific operating situation of the target bank operating institution, so that a more targeted analysis report can be generated by combining the report charts and the template text. Description of the Drawings

[0048] In order to more clearly illustrate the technical solutions in the embodiments of the present application or 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 some embodiments recorded in the present application. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0049] Figure 1 It is a schematic flowchart of an analysis report generation method provided by an embodiment of the present application;

[0050] Figure 2 It is a schematic diagram of the cover of an analysis report provided by an embodiment of the present application;

[0051] Figure 3 It is a schematic diagram of a system for an application analysis report generation method provided by an embodiment of the present application;

[0052] Figure 4 The structural schematic diagram of an analysis report generation device provided by an embodiment of the present application;

[0053] Figure 5 The structural schematic diagram of a computer device provided by an embodiment of the present application. Detailed implementation manners

[0054] Embodiments of the present application will be described in more detail below with reference to the accompanying drawings. Although some embodiments of the present application are shown in the drawings, it should be understood that the present application can be implemented in various forms and should not be construed as limited to the embodiments set forth herein. On the contrary, these embodiments are provided to more thoroughly and completely understand the present application. It should be understood that the drawings and embodiments of the present application are only for exemplary purposes and are not used to limit the protection scope of the present application.

[0055] As described in the background art, in order to reduce the manual writing cost of writing analysis reports, related technologies use automated analysis tools to generate analysis reports according to fixed templates. However, the automated analysis tools rely on pre-set fixed templates to output the text of the analysis report. When facing banking institutions with different scales, different business models or at different development stages, the analysis reports obtained by this "one-size-fits-all" method cannot fully reflect the unique attributes and specific operating conditions of each banking institution, thus lacking pertinence to banking institutions.

[0056] Based on this, embodiments of the present application provide an analysis report generation method and related device. Through a report text generation model, semantic understanding is performed on target basic data that can reflect the basic situation of the target banking institution itself and target operating data that can reflect the specific operating situation of the target banking institution, capturing the correlation and context relationships between the data, and converting the target basic data and target operating data into a first report text that can reflect the characteristics and specific operating situation of the target banking institution. Thus, a more targeted analysis report can be generated by combining report charts and template texts.

[0057] The analysis report generation method provided by the present application can be applied to computer devices with the ability to generate analysis reports, such as terminal devices and servers. Among them, the terminal device can specifically be a desktop computer, a laptop computer, a mobile phone, a tablet computer, etc.; the server can be an independent physical server, or a server cluster or distributed system composed of multiple physical servers, etc. The terminal device and the server can be directly or indirectly connected through wired or wireless communication methods, and the present application does not limit this here.

[0058] See Figure 1, This figure is a schematic flowchart of the analysis report generation method provided by the embodiments of the present application. For ease of description, in the following embodiments, the execution subject of this analysis report generation method is taken as an example of a server. As Figure 1 shown, this analysis report generation method includes S101 - S104.

[0059] S101: Obtain the template text, the target basic data of the target bank business institution, and the target business data of the target bank business institution.

[0060] A bank business institution is an independent organizational unit established by a banking financial institution, such as a branch, a sub - branch, a business department, a representative office, etc. The bank business institution is responsible for executing the financial business and services of the bank within a specific geographical area or business scope.

[0061] The target bank business institution is the bank business institution for which the analysis report is to be generated. The target basic data is the basic data of the target bank business institution, including the institution name, institution location, number of institution personnel, etc. The target business data is the business data of the target bank business institution, including deposits, loans, marketing activities, risk control, etc.

[0062] The template text is the fixed - format text of the analysis report, including texts with fixed - format requirements such as the cover, title, table of contents, chart positions, etc. The template text can be generated based on historical data and preset rules. For example, the template text may include the following content:

[0063] (1) Cover: As Figure 2 shown, the cover includes the report name (Analysis of the Business Development and Insights of Beijing XX Sub - branch), the report date, (December 1, 2024) the name of the report production unit.

[0064] (2) Title page: List the chapter titles of the analysis report, such as "I. Macroeconomic Environment", "II. Overview of Business Development", "III. Risk Management", etc.

[0065] (3) Table of contents.

[0066] (4) Reserved space for text and charts: Leave space in the corresponding chapter for inserting the first - hand report text and report charts.

[0067] (5) Standard paragraphs: Unified introduction text about the macro - economic situation.

[0068] It should be noted that all data collected in this application (such as target basic data or target business data) is collected with the consent and authorization of the data - belonging object (such as users, institutions, or enterprises), and the collection, use, and processing of relevant data need to comply with the relevant laws, regulations, and standards of relevant countries and regions.

[0069] S102: Based on the target basic data and the target business data, process them through the report text generation model to obtain the first report text of the target bank business institution.

[0070] The report text generation model is a large language model used to convert data into the first report text. A large language model (LLM) is a language model composed of an artificial neural network with many parameters (usually billions of weights or more), and is trained on a large amount of unlabeled text using self-supervised learning or semi-supervised learning. The first report text is a text for the business analysis of the bank business institution. This application does not specifically limit the report text generation model. For example, it can be obtained based on Transformer or Bidirectional Encoder Representations from Transformers (BERT). The first report text is a text for the analysis of the target business situation.

[0071] The embodiments of this application do not specifically limit the data processing method of the report text generation model. The following takes two methods as examples for illustration.

[0072] Method 1: Combine the target basic data and the target business data to obtain input data, input the input data into the report text generation model, and process it through the report text generation model to obtain the first report text. Specifically, first perform word segmentation on the input data to obtain multiple word segments, encode each word segment to obtain the word segment encoding vector corresponding to each word segment, set a separator vector between each word segment encoding vector to separate each word segment encoding vector, and then set the position vector corresponding to each word segment encoding vector to determine the order of each word segment encoding vector, so as to obtain the text encoding vector. Convert the text encoding vector through the model parameters of the report text generation model to obtain a token id sequence. Finally, perform decoding processing on the token id sequence to obtain the first report text.

[0073] Method 2: The report text generation model includes basic data and historical business data. The basic data includes the basic information data of bank business institutions of various institution types, and the historical business data includes the business data of each bank business institution that has been generated. Process the target basic data and the target business data through the report text generation model respectively. For details, refer to A1 - A3.

[0074] A1: Determine the institution category of the target bank business institution according to the target basic data and the basic data.

[0075] Among them, the institution category is the category divided according to the basic data of the bank business institution.

[0076] The report text generation model determines the institutional category of the target bank operating institution among various institutional categories by extracting features from the target basic data and matching the features with the basic data. For example, the basic data includes data of bank operating institutions of various personnel scales. The target basic data includes the number of institutional personnel. If the target bank operating institution has 10 institutional personnel, through feature comparison, the institutional category of the target bank operating institution is determined to be a small institution.

[0077] That is to say, by determining the institutional category of the target bank operating institution, it is equivalent to tagging the target bank operating structure. When conducting a specific analysis of the operating situation subsequently, the institutional category label will be referred to for specific analysis.

[0078] A2: Determine the similar operating data of bank operating institutions belonging to this institutional category from the historical operating data according to the institutional category of the target bank operating institution.

[0079] The similar operating data is the operating data of each bank operating institution belonging to the same institutional category as the target bank operating institution. By determining the similar operating data belonging to this institutional category from the historical operating data, the data range for analyzing the operating situation can be narrowed, and the similar operating data and the target operating data come from bank operating institutions belonging to the same institutional category. Therefore, when analyzing the operating situation of the target bank operating institution based on the similar operating data, the obtained analysis results are more targeted and accurate. For example, if the target bank operating institution is a small branch, the similar operating data is the operating data of all other small branches.

[0080] A3: Process according to the difference between the target operating data and the similar operating data through the report text generation model to obtain the first report text of the target bank operating institution.

[0081] Processing according to the difference between the target operating data and the similar operating data through the report text generation model can analyze the operating situation of the target bank operating institution among bank operating institutions of the same institutional type, so as to determine the specific operating situation corresponding to the operating data of the target bank operating institution among bank operating institutions of the same institutional type.

[0082] The operating data includes different types of data, and this difference is reflected in multiple aspects. For example, this difference can include financial indicators (such as net profit, operating income, cost control), business development (such as customer growth, market share change, new product launch situation), risk management (such as non-performing loan ratio, credit risk assessment), operational efficiency (such as per capita output, process automation degree), etc.

[0083] For example, if the loan growth rate of the target bank operating institution (Beijing XX Sub-branch) in 2024 is 10% higher than the average of similar small retail banks, the report text generation model can generate the text "Beijing XX Sub-branch achieved a loan growth significantly higher than the industry average in 2024".

[0084] By processing the differences between the target operating data and the similar operating data, the report text generation model outputs a complete first report text.

[0085] Thus, by determining the institution category of the target bank operating institution through the basic data, and performing tagging processing on the target operating data, and generating the first report text reflecting the operating conditions of the target bank operating institution through the differences between the target operating data and the similar operating data, the interference of cross-type operating data is avoided, and the accuracy of the analysis results is improved.

[0086] S103: Generate a report chart for the target bank operating institution according to the first report text.

[0087] The report chart is a visual chart generated for the first report text. To enhance the visualization effect of the analysis report, a chart generation tool with natural language processing capabilities can be used to generate multiple data points, so that the report chart can be drawn based on the multiple data points.

[0088] The following takes an example of a report chart generation method for illustration, see B1 - B3:

[0089] B1: Based on the first report text, identify through the chart generation model to obtain the target type of the report chart.

[0090] The chart generation model is a model with natural language processing capabilities and can generate charts. The target type is the type of the report chart, such as bar chart, line chart, pie chart, etc. Through semantic recognition of the first report text by the chart generation model, understanding the semantics expressed by the first report text, and according to the recognized results, determine the target type of the report chart. For example, if the first report text expresses that the deposit amount of the target bank operating institution has increased significantly compared to before in the time series, the target type of the report chart can be determined as a line chart.

[0091] B2: Determine the target generation method of the report chart according to the target type.

[0092] The target generation method is the specific implementation method for creating the report chart. Different target types correspond to different target generation methods. By according to the mapping relationship between the generation method and the type, the target generation method corresponding to the target type can be determined.

[0093] B3: Generate a report chart for the target bank operating institution according to the target generation method by means of the chart generation model.

[0094] For example, if it is determined according to the semantics expressed in the first report text that a line chart report chart should be generated, the chart generation model can be used to call the method function for generating a line chart to generate the report chart.

[0095] Thus, by performing semantic recognition on the first report text through the chart generation model, it is determined how to generate a chart that conforms to the corresponding content of the first report text, making the type of the report chart more matching with the content of the first report text, improving the accuracy of the report chart, and further enhancing the visualization effect of the analysis report.

[0096] S104: Generate an analysis report for the target bank operating institution according to the template text, the first report text, and the report chart.

[0097] As can be seen from the foregoing, the template text includes reserved positions for the first report text and the report chart. By inserting the first report text and the report chart into the corresponding positions, the template text, the first report text, and the report chart are merged to generate an analysis report for the target bank operating institution.

[0098] It can be seen from the above technical solution that the template text, the target basic data of the target bank operating institution, and the target operating data of the target bank operating institution are obtained. Based on the target basic data and the target operating data, the report text generation model is used for processing to obtain the first report text of the target bank operating institution. According to the first report text, a report chart of the target bank operating institution is generated. According to the template text, the first report text, and the report chart, an analysis report of the target bank operating institution is generated. Thus, through the report text generation model, semantic understanding is performed on the target basic data that can reflect the basic situation of the target bank operating institution itself and the target operating data that can reflect the specific operating situation of the target bank operating institution, the correlation relationship and the context relationship between the data are captured, and the target basic data and the target operating data are transformed into the first report text that can reflect the characteristics and specific operating situation of the target bank operating institution, so that a more targeted analysis report can be generated by combining the report chart and the template text.

[0099] In a possible implementation manner, the report text generation model is trained in the following manner. See C1-C3:

[0100] C1: Obtain target text sample pairs.

[0101] The target text sample pair includes a bank operating institution sample and the target report text corresponding to the bank operating institution sample, and is used to train the initial report text generation model. The bank operating institution sample includes the basic data and operating data of the bank operating institution. The target report text is a report text for the bank operating institution sample, and the target report text conforms to the language style and expression form of the bank analysis report and can be obtained by professional analysts writing the report.

[0102] C2: Based on the bank operating institution sample, use the initial report text generation model to make a prediction to obtain a predicted report text.

[0103] The initial report text generation model is used to predict the report text of the bank operating institution. The initial report text generation model is a report text generation model that has not been trained yet. It should be noted that the model structure of the initial report text model is the same as that of the report text generation model. When the training is completed, that is, when the model parameters of the initial report text generation model are adjusted, the report text generation model is obtained.

[0104] The predicted report text is the report text obtained by the initial report text generation model through prediction, and its accuracy may be low, and there may be differences from the language style and expression form of the bank analysis report.

[0105] C3: According to the differences between the predicted report text and the target report text, adjust the model parameters of the initial report text generation model to obtain the report text generation model.

[0106] The differences between the predicted report text and the target report text can reflect the accuracy of the initial report text generation model. Therefore, according to the differences between the predicted report text and the target report text, the model parameters of the initial report text generation model can be adjusted to make the differences smaller and smaller, thereby continuously improving the text generation accuracy of the initial report text generation model.

[0107] The embodiments of this application do not specifically limit the number and method of training. For example, after meeting the preset number of iterations or when the initial report text generation model converges, etc., the training of the initial report text generation model can be ended, that is, the model parameters of the initial report text generation model are no longer adjusted, so as to obtain a report text generation model with fixed parameters.

[0108] Thus, by training the initial report text generation model and continuously adjusting the model parameters to reduce the gap between the predicted report text and the target report text, the initial report text generation model learns how to more accurately capture the semantic information in the basic data and operating data and convert it into a report text with more accurate expression and style. Therefore, compared with other general text generation models, the report text generation model has a targeted improvement in generating the operating situation analysis text of bank operating institutions.

[0109] In a possible implementation, the embodiments of the present application can also generate an analysis report on the target bank operating institution according to external data, referring to D1 - D4:

[0110] D1: Obtain external data.

[0111] The external data includes external economic data and map data. For example, the external economic data includes industry trends, macro - economic indicators, etc., and the map data includes the geographical location of the target bank operating institution, the number of cooperative enterprises within a preset range, the number of public institutions, etc.

[0112] D2: Perform semantic recognition on the external data to obtain semantic information.

[0113] The semantic information is the result obtained by semantic recognition of the external data. Semantic recognition is to analyze the semantics of data or text through natural language processing technology, so as to extract semantic information with clear features from the external data, and further analyze the operating conditions of the target bank operating institution according to the semantic information. The embodiments of the present application do not limit the method of semantic recognition. The semantic recognition of the external data can be performed through a report text generation model, or can also be performed through other models with natural language processing capabilities.

[0114] D3: Based on the target basic data, target operating data, and semantic information, perform processing through a report text generation model to obtain a second report text of the target bank operating institution.

[0115] The second report text is a report text obtained by processing the external data through a report text generation model. The second report text can reflect the operating conditions of the target operating institution under the influence of external factors.

[0116] By combining the target basic data, target operating data, and semantic information and inputting them into the report text generation model, the report text generation model can fully understand the semantic information corresponding to the external data, so as to generate a second report text that can reflect the operating conditions of the target operating institution under the influence of external factors.

[0117] For example, if the external economic data indicates that the economic policy is favorable, while the turnover of the target bank operating institution shows a downward trend, the report text generation model will issue an abnormal warning in the second report text and indicate that further plans should be taken to adjust the operating conditions. Another example is that the location of the target bank operating institution is close to the urban area and the number of surrounding cooperative enterprises is large. The report text generation model will analyze the operating conditions in combination with the geographical location advantages of the target bank operating institution and output the analysis results in the second report text.

[0118] D4: Generate an analysis report for the target bank's business institution based on the template text, the first report text, the second report text, and the report charts.

[0119] The template text includes reserved positions for the first report text, the second report text, and the report charts. By inserting the first report text, the second report text, and the report charts into the corresponding positions, the template text, the first report text, the second report text, and the report charts are merged to generate an analysis report for the target bank's business institution.

[0120] Thus, by performing semantic recognition on external data and combining the target basic data and the target business data to generate the second report text, the analysis report can explain the business situation from the dimension of external factors, improving the comprehensiveness and accuracy of the analysis report.

[0121] In a possible implementation, the analysis report includes multiple chapters. The embodiments of the present application also provide a subscription mechanism for the analysis report. After generating the analysis report for the target bank's business institution, in response to obtaining a subscription request from a first user, determine the target chapter corresponding to the subscription request. Split the analysis report according to the target chapter to obtain the subscription content corresponding to the target chapter. Send the subscription content to the first user.

[0122] Among them, the first user is the user who issues the subscription request, and the first user can be one or more. For example, the first user can be an employee, a leader, etc. The subscription request is the request information or instruction for subscribing to the target chapter, the target chapter is the chapter of the analysis report that the first user hopes to view, and the subscription content is the content of the target chapter of the analysis report.

[0123] Thus, through this subscription mechanism, the content that the first user expects to see in the analysis report is sent to the first user, thereby meeting the needs of practitioners in different fields and enabling the analysis report to be sent more precisely to the subscribed first user.

[0124] In a possible implementation, the embodiments of the present application also provide a permission management mechanism. After generating the analysis report for the target bank's business institution, generate a permission identifier for the analysis report. In response to obtaining a distribution request from a second user, match the user represented by the permission identifier with the second user. If the second user belongs to the user represented by the permission identifier, distribute the analysis report to the second user. If the second user does not belong to the user represented by the permission identifier, do not perform distribution processing on the distribution request.

[0125] Among them, the permission identifier is used to represent the user who has the qualification to obtain the analysis report, the second user is the user who issues the distribution request, and the second user can be one or more. The distribution request is a request to distribute the analysis report to the second user.

[0126] Thus, through this permission management mechanism, it is ensured that the analysis report is only distributed to authorized users, that is, the users represented by the permission identifier, enhancing the data security and controllability of the analysis report, effectively preventing information leakage and maintaining data privacy.

[0127] See Figure 3 , Figure 3 which is a schematic diagram of a system for an application analysis report generation method provided by an embodiment of this application. The personnel economic database includes target basic data, the business indicator database includes target business data, the external economic database includes external economic data, and the map database includes map data. By inputting these four types of data into an algorithm library composed of three engines for processing, a first report text, a second report text, a template text, and report charts are obtained. Specifically, the conventional analysis report engine is used to generate the template text, and the visualization report engine is a chart generation model used to generate report charts for the first report text. Then, the analysis report is obtained through integration by the analysis report integration module program, and the analysis report is sent to the first user and the second user through the analysis report subscription and distribution program, where the user management program is used to manage permission information, subscription requirements, etc. Thus, through this system, an analysis report targeted at the target bank business institution can be generated.

[0128] See Figure 4 , Figure 4 which is an analysis report generation device provided by an embodiment of this application. The device 400 includes an acquisition unit 401, a processing unit 402, and a generation unit 403;

[0129] The acquisition unit 401 is configured to acquire a template text, the target basic data of the target bank business institution, and the target business data of the target bank business institution;

[0130] The processing unit 402 is configured to process based on the target basic data and the target business data through a report text generation model to obtain the first report text of the target bank business institution;

[0131] The generation unit 403 is configured to generate a report chart of the target bank business institution according to the first report text;

[0132] The generation unit 403 is further configured to generate an analysis report of the target bank business institution according to the template text, the first report text, and the report chart.

[0133] As can be seen from the above technical solution, the analysis report generation device provided in the embodiments of the present application includes an acquisition unit, a processing unit, and a generation unit. The acquisition unit acquires a template text, target basic data of the target bank operating institution, and target operating data of the target bank operating institution. The processing unit processes the target basic data and the target operating data through a report text generation model to obtain a first report text of the target bank operating institution. The generation unit generates a report chart of the target bank operating institution according to the first report text. The generation unit generates an analysis report of the target bank operating institution according to the template text, the first report text, and the report chart. Thus, through semantic understanding of the target basic data that can reflect the basic situation of the target bank operating institution itself and the target operating data that can reflect the specific operating situation of the target bank operating institution by the report text generation model, the correlation relationship and context relationship between the data are captured, and the target basic data and the target operating data are converted into a first report text that can reflect the characteristics and specific operating situation of the target bank operating institution, so that a more targeted analysis report can be generated by combining the report chart and the template text.

[0134] As a possible implementation manner, the report text generation model includes basic data and historical operating data, and the processing unit is specifically configured to:

[0135] Determine the institution category of the target bank operating institution according to the target basic data and the basic data;

[0136] According to the institution category of the target bank operating institution, determine the same type of operating data of the bank operating institutions belonging to the institution category from the historical operating data;

[0137] Process the difference between the target operating data and the same type of operating data through the report text generation model to obtain the first report text of the target bank operating institution.

[0138] As a possible implementation manner, the report text generation model is trained in the following manner:

[0139] Obtain a target text sample pair, where the target text sample pair includes a bank operating institution sample and a target report text corresponding to the bank operating institution sample;

[0140] Based on the bank operating institution sample, perform prediction through an initial report text generation model to obtain a predicted report text, where the initial report text generation model is used to predict the report text of the bank operating institution;

[0141] Adjust the model parameters of the initial report text generation model according to the differences between the predicted report text and the target report text to obtain the report text generation model.

[0142] As a possible implementation, the generating unit is specifically configured to:

[0143] Based on the first report text, identify through a chart generation model to obtain the target type of the report chart;

[0144] Determine the target generation method of the report chart according to the target type;

[0145] Generate the report chart of the target bank operating institution through the chart generation model according to the target generation method.

[0146] As a possible implementation, the device further includes an external analysis unit for:

[0147] Obtain external data, where the external data includes external economic data and map data;

[0148] Perform semantic recognition on the external data to obtain semantic information;

[0149] Based on the target basic data, the target operating data, and the semantic information, process through the report text generation model to obtain the second report text of the target bank operating institution;

[0150] Generate the analysis report of the target bank operating institution according to the template text, the first report text, the second report text, and the report chart.

[0151] As a possible implementation, the analysis report includes multiple chapters, and the device further includes a subscription unit for:

[0152] In response to obtaining a subscription request from a first user, determine the target chapter corresponding to the subscription request;

[0153] Split the analysis report according to the target chapter to obtain the subscription content corresponding to the target chapter;

[0154] Send the subscription content to the first user.

[0155] As a possible implementation, the device further includes a permission unit:

[0156] Generate a permission identifier for the analysis report, where the permission identifier is used to represent users who have the qualification to obtain the analysis report;

[0157] In response to obtaining a distribution request from a second user, match the user characterized by the permission identifier with the second user;

[0158] If the second user belongs to the user characterized by the permission identifier, distribute the analysis report to the second user.

[0159] See Figure 5 , an embodiment of the present application further provides a computer device, which includes a memory 501 and a processor 502:

[0160] The memory is used to store a computer program and transmit the computer program to the processor;

[0161] The processor is used to execute the method of the above method embodiment according to the computer program.

[0162] An embodiment of the present application further provides a computer-readable storage medium, which is characterized in that the computer-readable storage medium is used to store a computer program, and the computer program is used to execute the method of the above method embodiment.

[0163] An embodiment of the present application further provides a computer program product including a computer program, which, when running on a computer device, causes the computer device to execute the method of the above method embodiment.

[0164] It should be noted that the various embodiments in this specification are described in a progressive manner. Each embodiment focuses on the differences from other embodiments. The same or similar parts among the various embodiments can be referred to each other. For the systems or devices disclosed in the embodiments, since they correspond to the methods disclosed in the embodiments, the description is relatively simple, and the relevant parts can be referred to the description of the method part.

[0165] As used herein, the term "including" and its variants are open-ended, that is, "including but not limited to". The term "based on" means "at least partially based on". The term "one embodiment" means "at least one embodiment"; the term "another embodiment" means "at least one additional embodiment"; the term "some embodiments" means "at least some embodiments". The relevant definitions of other terms will be given in the following description.

[0166] It should be understood that in this application, "at least one (item)" means one or more, and "a plurality" means two or more. "And / or" is used to describe the association relationship of associated objects and indicates that there can be three relationships. For example, "A and / or B" can mean: only A exists, only B exists, and both A and B exist simultaneously. Here, A and B can be singular or plural. The character " / " generally indicates that the associated objects before and after are in an "or" relationship. "At least one (item) of the following" or similar expressions refer to any combination of these items, including any combination of single items or plural items. For example, at least one (item) of a, b, or c can mean: a, b, c, "a and b", "a and c", "b and c", or "a and b and c", where a, b, and c can be single or multiple.

[0167] It should also be noted that in this article, relational terms such as first and second are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the terms "comprise", "include" or any other variant thereof are intended to cover non-exclusive inclusion, such that a process, method, article or device comprising a series of elements not only includes those elements but also includes other elements not expressly listed, or elements inherent to such process, method, article or device. Without further limitation, an element defined by the statement "comprising an..." does not exclude the existence of additional identical elements in the process, method, article or device comprising the said element.

[0168] The steps of the methods or algorithms described in connection with the embodiments disclosed herein can be implemented directly in hardware, software modules executed by a processor, or a combination of both. The software modules can be placed in a random access memory (RAM), internal memory, read-only memory (ROM), electrically programmable ROM, electrically erasable programmable ROM, registers, hard disk, removable disk, CD-ROM, or any other form of storage medium well-known in the technical field.

[0169] The above description of the disclosed embodiments enables those skilled in the art to implement or use this application. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein can be implemented in other embodiments without departing from the spirit or scope of this application. Therefore, this application will not be limited to the embodiments shown herein, but rather to the broadest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for generating an analysis report, characterized in that: The method comprises: Obtaining a template text, target basic data of a target banking institution and target operating data of the target banking institution; Based on the target basic data and the target operating data, a report text generation model is used to process and obtain a first report text of the target banking institution; Generating a report chart of the target banking institution according to the first report text; An analysis report of the target banking institution is generated based on the template text, the first report text and the report chart.

2. The method according to claim 1, characterized in that The report text generation model includes basic data and historical operating data. The first report text of the target banking institution is obtained by processing the target basic data and the target operating data through the report text generation model, including: Determining the institution category of the target banking institution based on the target basic data and the basic data; According to the institution category of the target banking institution, determining similar operating data of banking institutions belonging to the institution category from the historical operating data; According to the difference between the target operating data and the similar operating data, the report text generation model is used to process the target operating institution to obtain a first report text.

3. The method according to claim 1, characterized in that The report text generation model is trained in the following way: Acquire a target text sample pair, the target text sample pair comprising a banking institution sample and a target report text corresponding to the banking institution sample; Based on the sample of banking institutions, prediction is performed by using an initial report text generation model to obtain a predicted report text, wherein the initial report text generation model is used to predict and obtain a report text of the banking institution; According to the difference between the predicted report text and the target report text, the model parameters of the initial report text generation model are adjusted to obtain the report text generation model.

4. The method according to claim 1, characterized in that Generating a report chart of the target banking institution according to the first report text includes: Based on the first report text, identifying by a chart generation model, obtaining a target type of the report chart; Determining a target generation method of the report chart according to the target type; The report chart of the target banking institution is generated by the chart generation model according to the target generation method.

5. The method according to claim 1, characterized in that: The method further comprises: Acquiring external data, wherein the external data includes external economic data and map data; Performing semantic recognition on the external data to obtain semantic information; Based on the target basic data, the target operating data and the semantic information, the report text generation model is used to process and obtain a second report text of the target banking institution; An analysis report of the target banking institution is generated based on the template text, the first report text, the second report text and the report chart.

6. The method according to claim 1, characterized in that The analysis report includes a plurality of chapters. After generating the analysis report of the target banking institution, the method further includes: In response to obtaining a subscription request from a first user, determining a target chapter corresponding to the subscription request; Splitting the analysis report according to the target chapters to obtain subscription content corresponding to the target chapters; The subscribed content is sent to the first user.

7. The method according to claim 1, characterized in that After generating the analysis report of the target banking institution, the method further includes: Generate an authorization identifier for the analysis report, where the authorization identifier is used to identify a user who is qualified to obtain the analysis report; In response to obtaining a distribution request from a second user, matching the user represented by the permission identifier with the second user; If the second user is a user represented by the permission identifier, the analysis report is distributed to the second user.

8. An analysis report generating device, characterized in that: The device comprises: an acquisition unit, a processing unit and a generation unit; The acquisition unit is used to acquire the template text, the target basic data of the target banking institution and the target operating data of the target banking institution; The processing unit is used to process the target basic data and the target operating data through a report text generation model to obtain a first report text of the target banking institution; The generating unit is used to generate a report chart of the target banking institution according to the first report text; The generating unit is further configured to generate an analysis report of the target banking institution based on the template text, the first report text and the report chart.

9. A computer device, characterized in that: The computer device comprises a processor and a memory: The memory is used to store a computer program and transmit the computer program to the processor; The processor is configured to execute the method according to any one of claims 1 to 7 according to the computer program.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium is used to store a computer program, and the computer program is used to execute the method according to any one of claims 1 to 7.

Images