Method and system for analyzing financial data

Abstract

The embodiment of the application provides a kind of financial data analysis method and system.The application is applied to intelligent agent platform, and analysis method includes: receiving the business analysis text input by user;Business object, business time range and business analysis type are obtained by parsing business analysis text, and analysis task is generated;Analysis request is initiated to BI platform;The result data set returned by BI platform for analysis request is received, and result data set includes business object and business index, and business index includes business data, abnormal data and the abnormal amplitude of abnormal data, the dimension where it is and abnormal reason;Result data set is obtained after BI platform executes data query and abnormal data analysis to analysis task, and data query refers to that BI platform carries out semantic mapping based on metadata to obtain business data and carries out data calculation, and abnormal data analysis refers to that BI platform carries out abnormal detection to the business data obtained after executing data query based on abnormal data model;Analysis result text is generated according to result data set.

Description

Technical Field

[0001] This application relates to the field of data analysis technology, specifically to a method and system for analyzing financial data. Background Technology

[0002] In the field of financial management, anomaly analysis of financial statement data is a crucial step in risk identification and business decision-making. Currently, the industry primarily relies on the following methods to detect and assess anomalies: One approach is purely manual analysis. Experienced financial professionals manually compare multi-dimensional data and cross-validate the relationships between various business reports to identify data fluctuations or abnormal trends. This method heavily relies on individual expertise and experience; different analysts may arrive at different conclusions, making consistency difficult to guarantee. Furthermore, manual operation is limited by processing power and is typically only suitable for offline scenarios with small datasets. When faced with massive amounts of financial data, manual analysis efficiency drops drastically. It is not only time-consuming and labor-intensive but also struggles to achieve comprehensive coverage and real-time response, often relying on sampling or post-event review, failing to promptly identify risks.

[0003] Another approach involves leveraging big data query and configuration tools for auxiliary analysis. This method typically requires finance personnel to submit analysis requests, followed by big data developers writing complex query scripts, configuring data mining rules, or building specific analytical models to explore relationships between data and filter out anomalies. This "business requests, technology implementation" collaborative model incurs significant communication costs and technical barriers. Implementing an analysis request often requires repeated confirmation of definitions and adjustments to logic, resulting in a lengthy response cycle. Furthermore, the quality of the analysis results is limited by both the business abstraction capabilities of finance personnel and the technical implementation capabilities of developers. A misunderstanding or lack of experience at either stage can affect the accuracy of the final conclusions, making it difficult to provide effective support for business decisions.

[0004] Furthermore, most existing software products or platforms on the market lack direct intelligent anomaly analysis capabilities. Their typical approach involves displaying potential anomalies in charts or lists using pre-set business logic rules or big data mining algorithms, ultimately requiring financial personnel to rely on experience for manual judgment and characterization. Throughout this process, the initiative in anomaly detection and the depth of attribution analysis are insufficient. Meanwhile, current corporate financial statement data is generally characterized by high update frequency and constantly changing business needs. Traditional analysis methods based on fixed rule configurations require technical personnel to manually adjust the underlying configuration and query logic when business rules or data definitions change. This static rule mechanism is ill-suited to adapting to rapidly iterating business scenarios and diverse analytical demands, resulting in high maintenance costs and insufficient timeliness. Summary of the Invention

[0005] The purpose of this application is to provide a method and system for analyzing financial data, in order to solve the technical defects in the prior art where the configuration of analysis rules that rely on human experience cannot meet the needs of rapidly iterating business analysis.

[0006] To achieve the above objectives, the first aspect of this application provides a method for analyzing financial data, applied to an intelligent agent platform, wherein the intelligent agent platform interacts with a BI platform, and the analysis method includes: Receive business analysis text input by the user; Parse the business analysis text to obtain the business object, business time range, and business analysis type, and generate the analysis task; An analysis request is sent to the BI platform based on the analysis task. The BI platform is configured with metadata and anomaly data models corresponding to the user's required business model. The system receives the result dataset returned by the BI platform in response to the analysis request. The result dataset includes business objects and business metrics. The business metrics include at least business data, abnormal data, and the magnitude, dimension, and cause of the abnormal data. The result dataset is obtained by the BI platform after performing data query and abnormal data analysis on the analysis task in sequence. Data query refers to the BI platform obtaining business data and performing data calculations based on metadata semantic mapping. Abnormal data analysis refers to the BI platform performing anomaly detection on the business data obtained after executing the data query based on the abnormal data model. Generate and output analysis result text based on the result dataset.

[0007] In the embodiments of this application, parsing business analysis text to obtain business objects, business time ranges, and business analysis types, and generating analysis tasks includes: parsing business analysis text to obtain business objects, business time ranges, and business analysis types, and determining whether the business analysis text is the user's first business analysis text; if the business analysis text is the first business analysis text, generating an analysis task; if the business analysis text is not the first business analysis text, retrieving the user's historical business analysis texts, identifying the historical business objects of the historical business analysis texts as business objects, and generating analysis tasks.

[0008] In the embodiments of this application, the intelligent agent platform is connected to the client, and the analysis method further includes: generating an abnormal data alert when there is abnormal data in the result dataset, and pushing the abnormal data alert to the client.

[0009] A second aspect of this application provides a method for analyzing financial data, applied to a BI platform, wherein the BI platform interacts with an intelligent agent platform and an indicator platform, and the analysis method includes: The system receives metadata sent by the metrics platform and builds an anomaly data model based on the metadata. The metadata is extracted by the metrics platform from the business model after the user builds the business model based on the data definition component provided by the metrics platform according to the required business scenario. In response to the analysis request sent by the intelligent agent platform, the analysis task is semantically mapped based on metadata to retrieve business data, and data calculations are performed on the business data. Anomaly detection is performed on the calculated business data based on the anomaly data model, and a result dataset is generated for the business data after anomaly detection. The result dataset includes business objects and business metrics. The business metrics include at least business data, abnormal data, and the magnitude, dimension, and cause of the anomaly. The resulting dataset is returned to the agent platform.

[0010] In the embodiments of this application, the abnormal data model performs anomaly detection on the calculated business data and generates a result dataset for the anomaly-detected business data, including: determining whether the calculated business data is abnormal based on the abnormal data model; if the calculated business data is abnormal, determining the dimension, magnitude, and cause of the abnormal business data in the database table, and generating a result dataset based on the abnormal business data and the magnitude, dimension, and cause of the abnormal business data; if the business data is not abnormal, generating a result dataset based on normal business data.

[0011] In the embodiments of this application, metadata includes at least one of entities, fields, relationships, metric definitions, and permission configurations. The BI platform maps the metadata in the metric platform to metadata that conforms to the user's business behavior.

[0012] A third aspect of this application provides a financial data analysis system, the analysis system comprising: The intelligent agent platform receives business analysis text input by the user; parses the business analysis text to obtain the business object, business time range, and business analysis type, and generates an analysis task; based on the analysis task, it sends an analysis request to the BI platform, where the BI platform is configured with metadata and abnormal data models corresponding to the user's required business model; it receives the result dataset returned by the BI platform in response to the analysis request, where the result dataset includes business objects and business metrics, and the business metrics include at least business data, abnormal data, and the magnitude, dimension, and cause of the abnormality of the abnormal data; and it generates and outputs the analysis result text based on the result dataset. The BI platform receives metadata sent by the metrics platform and builds anomaly data models based on the metadata; responds to analysis requests sent by the intelligent agent platform, performs semantic mapping on the analysis tasks based on the metadata to retrieve business data, and performs data calculations on the business data; performs anomaly detection on the calculated business data based on the anomaly data models, and generates a result dataset for the anomaly-detected business data; and returns the result dataset to the intelligent agent platform.

[0013] In embodiments of this application, the analysis system further includes: an indicator platform, used to provide a data definition component to extract metadata from the business model after the user builds a business model based on the data definition component according to the required business scenario; and to send the metadata to the BI platform.

[0014] In the embodiments of this application, the indicator platform is also used to: receive the core indicator calculation caliber and the abnormal data threshold of the calculation caliber from user-defined business data; and associate the core indicator calculation caliber and the abnormal data threshold of the calculation caliber with the corresponding business model.

[0015] In embodiments of this application, the metrics platform is also used to: after extracting metadata from the business model, convert the metadata into a standardized metadata format that the BI platform can recognize and unify the semantic tags of the fields.

[0016] The above technical solution involves receiving business analysis text input by the user through an intelligent agent platform; parsing the business analysis text to obtain the business object, business time range, and business analysis type, and generating an analysis task; initiating an analysis request to the BI platform based on the analysis task, wherein the BI platform is configured with metadata and anomaly data models corresponding to the user's required business model; receiving the result dataset returned by the BI platform in response to the analysis request, wherein the result dataset includes business objects and business indicators, and the business indicators include at least business data, anomaly data, and the anomaly magnitude, dimension, and cause of the anomaly; the result dataset is obtained by the BI platform after sequentially performing data query and anomaly data analysis on the analysis task, where data query refers to the BI platform obtaining business data and performing data calculations based on metadata semantic mapping, and anomaly data analysis refers to the BI platform performing anomaly detection on the business data obtained after executing data query based on the anomaly data model; and generating and outputting analysis result text based on the result dataset. This invention, through interaction with an intelligent agent platform, enables rapid iteration of business analysis needs, allowing business personnel to independently complete anomaly analysis without technical expertise, and enabling financial personnel to independently perform automated and efficient anomaly financial data analysis.

[0017] Other features and advantages of the embodiments of this application will be described in detail in the following detailed description section. Attached Figure Description

[0018] The accompanying drawings are provided to further illustrate the embodiments of this application and form part of the specification. They are used together with the following detailed description to explain the embodiments of this application, but do not constitute a limitation on the embodiments of this application. In the drawings: Figure 1 This illustration schematically shows an application environment diagram of a financial data analysis method according to an embodiment of this application; Figure 2 The illustration shows a flowchart of a financial data analysis method according to an embodiment of this application; Figure 3 The schematic diagram illustrates a flowchart of another financial data analysis method according to an embodiment of this application; Figure 4 The diagram illustrates the internal structure of a computer device according to an embodiment of this application. Detailed Implementation

[0019] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. It should be understood that the specific embodiments described herein are only for illustration and explanation of the embodiments of this application and are not intended to limit the embodiments of this application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without creative effort are within the scope of protection of this application.

[0020] Furthermore, if the embodiments of this application involve descriptions such as "first" or "second," these descriptions are for descriptive purposes only and should not be construed as indicating or implying their relative importance or implicitly specifying the number of technical features indicated. Therefore, features defined with "first" or "second" may explicitly or implicitly include at least one of those features. Additionally, the technical solutions of various embodiments can be combined with each other, but this must be based on the ability of those skilled in the art to implement them. If the combination of technical solutions is contradictory or impossible to implement, it should be considered that such a combination of technical solutions does not exist and is not within the scope of protection claimed in this application.

[0021] This application provides a method for analyzing financial data, which can be applied to, for example... Figure 1 The application environment shown. Among them, such as... Figure 1As shown, the intelligent agent platform can refer to an artificial intelligence agent platform. Users, such as finance personnel, can use the indicator platform to build business models and interactive interfaces with zero code. After the business model is built, the indicator platform's automatic metadata synchronization mechanism enables seamless integration between the low-code business model and the BI platform. Relying on the artificial intelligence agent platform, a natural language interaction entry point is achieved. Finally, through the collaborative analysis of the large model and the BI platform, the intelligent agent platform can achieve intelligent identification, deep attribution, and visualization of abnormal data.

[0022] Figure 2 The illustration schematically shows a flowchart of a financial data analysis method according to an embodiment of this application. Figure 2 As shown in the illustration, this application provides a method for analyzing financial data, applied to an intelligent agent platform. The intelligent agent platform interacts with a BI platform. This analysis method may include the following steps: Step 201: Receive the business analysis text input by the user.

[0023] In this embodiment, it should be noted that the business analysis text can refer to the text content entered by the user in natural language on the interactive interface of the intelligent agent platform according to actual business needs. The entered business analysis text does not need to follow a preset instruction template or a fixed keyword format, but can directly express the analysis request using the natural language habits of daily communication. Taking financial personnel as an example, they can enter the reason for the decline in sales revenue in a certain month of a certain year on the interactive interface of the intelligent agent platform, such as: "Please analyze whether there are any abnormalities in this month's sales revenue? What are the reasons for this?"

[0024] Step 202: Parse the business analysis text to obtain the business object, business time range, and business analysis type, and generate the analysis task.

[0025] In this embodiment, it should be noted that after receiving business analysis text input by the user, the intelligent agent platform can use an NLU (Natural Language Understanding) engine to perform intent recognition and demand parsing on the text, extracting key information elements such as the business object, business time range, and business analysis type. Then, it outputs the corresponding analysis task through a pre-configured analysis module. For example, if a finance staff member inputs: "Help me analyze whether there are any anomalies in this month's sales revenue? What could be the cause?", the intelligent agent platform can use the NLU engine to parse the business object as sales revenue, the business time range as this month, and the business analysis type as anomaly detection and attribution analysis. Based on the above parsing results, the intelligent agent platform generates a structured analysis task containing the above three elements, which is used to initiate subsequent analysis requests to the BI platform.

[0026] In this embodiment of the application, parsing the business analysis text to obtain the business object, business time range, and business analysis type, and generating the analysis task includes: parsing the business analysis text to obtain the business object, business time range, and business analysis type, and determining whether the business analysis text is the user's first business analysis text; if the business analysis text is the first business analysis text, generating the analysis task; if the business analysis text is not the first business analysis text, retrieving the user's historical business analysis texts, identifying the historical business objects of the historical business analysis texts as the business objects, and generating the analysis task.

[0027] In this embodiment, it should be noted that before generating an analysis task, the intelligent agent platform can determine whether the business analysis text is the user's first business analysis text. The AI ​​agent platform automatically records historical conversations with users; therefore, it can determine this by searching the current user's historical conversation records to see if there are any analysis task records associated with that user. Specifically, if the business analysis text is the first business analysis text (i.e., the intelligent agent platform has not found any historical analysis task records associated with the current user), the intelligent agent platform can directly generate an analysis task based on the currently parsed business object, business time range, and business analysis type. If the business analysis text is not the first business analysis text (i.e., the intelligent agent platform has found historical analysis task records associated with the current user), the intelligent agent platform can extract historical business objects from the historical business analysis text, use these historical business objects as the business object of the current analysis task, and combine them with the business time range and business analysis type parsed from the current text to jointly generate the analysis task. For example, if a finance employee initially enters a message asking for an analysis of this month's sales revenue for any anomalies, the AI ​​agent platform parses the data to identify the business object as sales revenue, the time frame as this month, and the analysis type as anomaly detection and attribution analysis. It then determines this is the first analysis request and directly generates an analysis task. When the finance employee continues by entering information about last month, the AI ​​agent platform can retrieve the previously identified business object, sales revenue, from historical conversations. Therefore, it can use this as the business object for the current analysis task and generate a new analysis task accordingly.

[0028] Step 203: Initiate an analysis request to the BI platform based on the analysis task. The BI platform is configured with metadata and anomaly data models corresponding to the business model required by the user.

[0029] In this embodiment, it should be noted that after generating an analysis task, the intelligent agent platform can initiate an analysis request to the BI platform based on the analysis task. This analysis request includes information such as the business object, the business time range, and the business analysis type. The BI platform is pre-configured with metadata corresponding to the user's required business model. This metadata may include entities, fields, relationships, indicator definitions, and permission configurations. Simultaneously, the BI platform is also pre-configured with an anomaly data model, which is used to detect and judge anomalies in business data in subsequent steps.

[0030] Step 204: Receive the result dataset returned by the BI platform in response to the analysis request. The result dataset includes business objects and business metrics. The business metrics include at least business data, abnormal data, and the magnitude, dimension, and cause of the abnormality of the abnormal data. The result dataset is obtained by the BI platform after performing data query and abnormal data analysis on the analysis task in sequence. Data query refers to the BI platform obtaining business data and performing data calculations based on metadata semantic mapping. Abnormal data analysis refers to the BI platform performing anomaly detection on the business data obtained after executing the data query based on the abnormal data model.

[0031] In this embodiment, it should be noted that after the intelligent agent platform initiates an analysis request to the BI platform, the BI platform responds to the analysis request and performs analysis, obtaining a result dataset, which is then returned to the intelligent agent platform. The result dataset may contain business objects and corresponding business metrics. Business metrics may include business data, abnormal data, and the magnitude, dimension, and cause of the abnormality. Business data may refer to actual operating data obtained by the BI platform after semantic mapping and data calculation based on metadata, such as this month's sales revenue or the sales revenue of the same period last year. Abnormal data may refer to data points identified as abnormal by the BI platform after anomaly detection based on an abnormal data model. The magnitude of the abnormality may refer to the degree of change of the abnormal data compared to the baseline data, such as a year-on-year decrease percentage. The dimension may refer to the main contributing dimension located after multi-dimensional attribution analysis, such as overseas regions in the regional dimension or mobile phone categories in the product dimension. The cause of the abnormality may refer to the attribution conclusions drawn after analyzing the contribution of the abnormal data across various dimensions.

[0032] Step 205: Generate and output the analysis result text based on the result dataset.

[0033] In this embodiment, it should be noted that after receiving the result dataset returned by the BI platform, the intelligent agent platform can generate analysis result text based on the result dataset and output the analysis result text to the interactive interface for user viewing. The analysis result text may include a basic description of the business object, anomaly judgment conclusions, anomaly magnitude descriptions, and attribution analysis results. The analysis result text can be generated by filling the structured data in the result dataset into a preset natural language template, or by the intelligent agent platform organizing the result dataset into natural language. Taking a finance staff member querying anomalies in this month's sales revenue as an example, the intelligent agent platform receives a result dataset containing data such as the business object being sales revenue, the anomaly magnitude being a year-on-year decrease of 15%, the anomaly dimension being overseas regions, and the cause of the anomaly. Based on this, the generated analysis result text could be: This month's sales revenue is abnormal, decreasing by 15% year-on-year, mainly due to a decline in sales in overseas regions, with overseas regions contributing 80% of this decrease. This analysis result text is output to the intelligent agent platform's interactive interface, allowing users to directly obtain the analysis conclusion without having to interpret the structured data themselves.

[0034] In this embodiment of the application, the intelligent agent platform is connected to the client, and the analysis method further includes: generating an abnormal data alert when there is abnormal data in the result dataset, and pushing the abnormal data alert to the client.

[0035] In this embodiment, it should be noted that the intelligent agent platform can also have an abnormal data push function. The intelligent agent platform can communicate with the client and push abnormal data to the client in a timely manner so that the user can view the information promptly. The client can refer to the interactive interface on the user's terminal device or a third-party instant messaging platform. Specifically, after receiving the result dataset returned by the BI platform, the intelligent agent platform can judge the result dataset. If there is abnormal data in the result dataset, it generates an abnormal data alert and pushes the abnormal data alert to the client. The abnormal data alert can include key information such as the business object, the magnitude of the abnormality, and the dimension in which the abnormality occurs. Taking the example of a finance staff member querying abnormal sales revenue for the current month, if there is abnormal data in the result dataset received by the intelligent agent platform, it will generate an abnormal data alert. The content may include that the sales revenue for the current month is abnormal, with a year-on-year decrease of 15%, mainly attributed to overseas regions. This alert will be pushed to the finance staff member's instant messaging tool for timely viewing.

[0036] The above technical solution involves receiving business analysis text input by the user through an intelligent agent platform; parsing the business analysis text to obtain the business object, business time range, and business analysis type, and generating an analysis task; initiating an analysis request to the BI platform based on the analysis task, wherein the BI platform is configured with metadata and anomaly data models corresponding to the user's required business model; receiving the result dataset returned by the BI platform in response to the analysis request, wherein the result dataset includes business objects and business indicators, and the business indicators include at least business data, anomaly data, and the anomaly magnitude, dimension, and cause of the anomaly; the result dataset is obtained by the BI platform after sequentially performing data query and anomaly data analysis on the analysis task, where data query refers to the BI platform obtaining business data and performing data calculations based on metadata semantic mapping, and anomaly data analysis refers to the BI platform performing anomaly detection on the business data obtained after executing data query based on the anomaly data model; and generating and outputting analysis result text based on the result dataset. This invention, through interaction with an intelligent agent platform, enables rapid iteration of business analysis needs, allowing business personnel to independently complete anomaly analysis without technical expertise, and enabling financial personnel to independently perform automated and efficient anomaly financial data analysis.

[0037] Figure 3 A schematic flowchart illustrating another method for analyzing financial data according to an embodiment of this application is shown. Figure 3 As shown in the embodiment of this application, a method for analyzing financial data is provided and applied to a BI platform. The BI platform interacts with an intelligent agent platform and an indicator platform. The analysis method may include the following steps: Step 301: Receive metadata sent by the metrics platform and build an abnormal data model based on the metadata. The metadata is extracted by the metrics platform from the business model after the user builds the business model based on the data definition component provided by the metrics platform according to the required business scenario.

[0038] In this embodiment, it should be noted that the BI platform can receive metadata sent by the metrics platform. This received metadata can refer to data automatically extracted and synchronized to the BI platform by the metrics platform after the user constructs a business model based on their specific business scenario using the data definition component provided by the metrics platform. Specifically, taking a finance professional building a sales revenue model as an example, the finance professional can use the data definition component of the metrics platform to customize and construct a sales revenue business model containing order entities, customer entities, and product entities. After the business model is configured, the metrics platform automatically extracts the model's metadata and synchronizes it to the BI platform. After receiving the metadata, the BI platform can construct a corresponding abnormal data model based on the metadata, providing a foundation for performing anomaly detection when receiving analysis requests from the intelligent agent platform.

[0039] In this embodiment of the application, metadata includes at least one of entities, fields, relationships, metric definitions, and permission configurations. The BI platform maps the metadata in the metric platform to metadata that conforms to the user's business behavior.

[0040] In this embodiment, it should be noted that metadata may include entity definitions, field definitions, relationships between entities, metric calculation methods, and permission configurations. Taking an entity as an example, an entity can refer to a core business object in a business scenario, such as an order entity, a user entity, and a host entity. Taking a field as an example, a field can refer to a specific attribute under an entity, such as order ID, order time, sales amount, order status, and installment method under an order entity; user ID, region, and country of origin under a user entity; and host serial number ID, product category, supplier, and inventory quantity under a host entity. Taking a relationship as an example, a relationship can refer to the connection established between entities through shared fields, such as an order entity and a customer entity being associated through user ID, and an order and a product being associated through host serial number ID. Taking indicator definitions as an example, indicator definitions can refer to user-defined indicator calculation logic, which may include core indicator calculation definitions and thresholds for abnormal data in the calculation definitions. For example, repurchase rate is defined as the percentage of users who have purchased twice or more in the past 30 days, and sales year-on-year growth rate is defined as the difference between current sales and sales in the same period last year / sales in the same period last year. For example, when sales fluctuate by more than ±20%, a red warning is automatically marked in the data display. Taking permission configuration as an example, permission configuration can refer to the association and binding relationship between business models and dimension sets, used to control the relevant indicator permissions of different roles through the association results. For user use, the BI platform maps the metadata in the indicator platform to metadata that conforms to the user's business behavior. For example, the sale_order_amt field in the database is mapped to the business term "sales revenue," and the data_source field is mapped to "data source," unifying the indicator definitions from the root and avoiding the problem of inconsistent definitions between different departments.

[0041] It should be noted that since metadata includes entities, fields, relationships, metric definitions, and permission configurations, the abnormal data model built based on metadata can include threshold rules for anomaly judgment, fluctuation detection logic, etc., providing a foundation for performing anomaly detection when receiving analysis requests from the intelligent agent platform.

[0042] Step 302: In response to the analysis request sent by the intelligent agent platform, perform semantic mapping on the analysis task based on metadata to retrieve business data, and perform data calculation on the business data.

[0043] In this embodiment, it should be noted that after receiving an analysis request from the intelligent agent platform, the BI platform can respond to the request by performing semantic mapping on the analysis task based on pre-configured metadata. Semantic mapping refers to mapping business terms such as business objects and time ranges in the analysis task to corresponding technical fields and query conditions in the underlying data source. Based on the result of semantic mapping, the BI platform can initiate a data query request to the underlying data source to retrieve the corresponding raw business data. After obtaining the raw business data, the BI platform can also perform data calculations on the raw business data according to the indicator definitions in the metadata to obtain the business data required for the analysis task.

[0044] Specifically, taking an example of a finance staff member querying for anomalies in this month's sales revenue, the analysis request sent by the intelligent agent platform includes the sales revenue of the business object and the business time range for this month. Based on metadata, the BI platform maps sales revenue to sales amount-related fields in the underlying data source, and maps this month to the corresponding start and end date range. Based on this, it initiates a query to the underlying data source to retrieve the sales details for this month. Simultaneously, the BI platform summarizes and calculates the retrieved sales details data according to the indicator definitions in the metadata to obtain the sales revenue amount for this month. Furthermore, if the analysis task involves year-on-year and month-on-month comparisons, the BI platform can also simultaneously calculate the sales revenue amount for the same period last year as benchmark data for subsequent anomaly detection.

[0045] Step 303: Based on the abnormal data model, perform anomaly detection on the calculated business data and generate a result dataset for the business data after anomaly detection. The result dataset includes business objects and business metrics. The business metrics include at least business data, abnormal data, and the magnitude, dimension, and cause of the anomaly.

[0046] In this embodiment, it should be noted that after the BI platform completes data calculation to obtain business data, it can perform anomaly detection on the calculated business data based on a pre-built anomaly data model. Anomaly detection refers to comparing the business data with preset benchmark data or threshold rules in the anomaly data model to determine whether the business data exhibits abnormal fluctuations. If an anomaly is determined, the BI platform can further perform multi-dimensional attribution analysis on the anomaly data to determine the contribution of the anomaly data in each dimension, thereby locating the dimension where the anomaly is located and deriving the cause of the anomaly.

[0047] After anomaly detection and attribution analysis are completed, the BI platform can generate a results dataset. The results dataset can include business objects and business metrics. Business metrics must include at least business data, anomalous data, and the magnitude, dimension, and cause of the anomaly. Specifically, business data refers to calculated actual operational data; anomalous data refers to the data identified as anomalous; the magnitude of the anomaly refers to the degree of change of the anomalous data compared to the baseline data; the dimension refers to the main contributing dimension identified after attribution analysis; and the cause of the anomaly refers to the attribution conclusions drawn after analyzing the contribution of each dimension.

[0048] Specifically, taking an anomaly in monthly sales revenue as an example, the BI platform, based on an anomaly data model, compares the calculated monthly sales revenue with the sales revenue of the same period last year. It finds that the year-on-year decline exceeds a preset threshold, thus identifying it as an anomaly. Secondly, the BI platform performs multi-dimensional drill-down attribution analysis on the anomaly data, pinpointing the anomaly to the overseas region. The cause of the anomaly is that the decline in sales in the overseas region contributed the most to the overall revenue decline. Finally, in the result dataset generated by the BI platform, the business object is sales revenue, the business indicators can include the monthly sales revenue amount and the sales revenue amount of the same period last year, the anomaly data is the monthly sales revenue data, the anomaly magnitude is the specific percentage of year-on-year decline, the dimension is the overseas region, and the cause of the anomaly is an explanation of the specific proportion of decline contributed by the overseas region.

[0049] In this embodiment of the application, the anomaly detection of the calculated business data based on the anomaly data model and the generation of a result dataset for the anomaly-detected business data include: determining whether the calculated business data is abnormal based on the anomaly data model; if the calculated business data is abnormal, determining the dimension, magnitude, and cause of the abnormal business data in the database table, and generating a result dataset based on the abnormal business data and the magnitude, dimension, and cause of the abnormal business data; if the business data is not abnormal, generating a result dataset based on normal business data.

[0050] In this embodiment, it should be noted that when the BI platform performs anomaly detection on the calculated business data based on the anomaly data model, it can first determine whether the calculated business data is abnormal. For example, it can compare the business data with the preset benchmark data in the anomaly data model, calculate the change range, and determine whether the change range exceeds the preset threshold.

[0051] Specifically, if anomalies are found in the calculated business data, the BI platform further analyzes the dimension, magnitude, and cause of the anomaly in the database table. The dimension can refer to the main contributing dimension identified through multi-dimensional drill-down analysis; the magnitude of the anomaly refers to the degree of change of the anomaly data compared to the baseline data; and the cause of the anomaly refers to the attribution conclusions drawn after analyzing the contribution of each dimension. After obtaining the dimension, magnitude, and cause of the anomaly, a result dataset is generated. If the business data does not exhibit anomalies, i.e., the magnitude of change does not exceed a preset threshold, a result dataset is generated based on normal business data. This result dataset can contain business objects and normal business data, and is marked as indicating that no anomalies were detected.

[0052] Specifically, taking a finance professional's inquiry into this month's sales revenue as an example, the BI platform compares this month's sales revenue with the same period last year. If the year-on-year decline exceeds a preset threshold, it is considered an anomaly. Then, further drill-down analysis is performed at the regional and product levels to pinpoint the anomaly as the overseas region, with an anomaly of 15% year-on-year decline. The cause of the anomaly is that the decline in overseas sales contributed 80% to the overall decline. The resulting dataset contains the abnormal business data, the magnitude of the anomaly, the region where the anomaly occurred, and the cause of the anomaly. Conversely, if this month's sales revenue fluctuates within the normal range compared to the same period last year, it is considered normal, and the generated dataset contains normal business data and a marker indicating no anomalies were detected.

[0053] Step 304: Return the result dataset to the agent platform.

[0054] In this embodiment of the application, it should be noted that after the BI platform generates a result dataset through a series of analysis and processing, it responds to the analysis request of the intelligent agent platform and returns the result dataset to the intelligent agent platform so that the intelligent agent platform can subsequently output the result dataset to the user.

[0055] This application provides a financial data analysis system, comprising: an intelligent agent platform for receiving business analysis text input by a user; parsing the business analysis text to obtain business objects, business time ranges, and business analysis types, and generating analysis tasks; initiating analysis requests to a BI platform based on the analysis tasks, wherein the BI platform is configured with metadata and anomaly data models corresponding to the business model required by the user; receiving a result dataset returned by the BI platform in response to the analysis request, wherein the result dataset includes business objects and business indicators, and the business indicators include at least business data, anomaly data, and the anomaly magnitude, dimension, and cause of the anomaly data; generating and outputting analysis result text based on the result dataset; and a BI platform for receiving metadata sent by an indicator platform and constructing anomaly data models based on the metadata; responding to the received analysis requests sent by the intelligent agent platform, performing semantic mapping on the analysis tasks based on the metadata to retrieve business data, and performing data calculations on the business data; performing anomaly detection on the calculated business data based on the anomaly data models, and generating a result dataset for the anomaly-detected business data; and returning the result dataset to the intelligent agent platform.

[0056] In this embodiment, it should be noted that the user inputs business analysis text in natural language through the interactive interface of the intelligent agent platform. After receiving the business analysis text, the intelligent agent platform parses the text using an NLU engine to obtain the business object, business time range, and business analysis type, and generates an analysis task. The intelligent agent platform then sends an analysis request to the BI platform. The BI platform stores metadata sent by the metrics platform and an anomaly data model built based on the metadata. Upon receiving the analysis request, the BI platform performs semantic mapping on the analysis task based on the metadata, calls the underlying data source to obtain the corresponding raw business data and performs data calculations, then performs anomaly detection on the calculated business data based on the anomaly data model, generates a result dataset, and returns it to the intelligent agent platform. The result dataset includes business objects and business metrics. Business metrics include business data, anomaly data, anomaly magnitude, dimension, and cause of the anomaly. After receiving the result dataset, the intelligent agent platform generates analysis result text based on the result dataset and outputs it to the interactive interface.

[0057] In this embodiment of the application, the analysis system further includes: an indicator platform, which provides a data definition component to extract metadata from the business model after the user builds a business model based on the data definition component according to the required business scenario; and sends the metadata to the BI platform.

[0058] In this embodiment, it should be noted that the metrics platform can provide a data definition component, which allows users to build business models according to their required business scenarios. After the business model is built, the metrics platform can automatically extract metadata from the business model and send the extracted metadata to the BI platform. The metadata may include entities, fields, relationships, metric definitions, and permission configurations, etc.

[0059] In this embodiment of the application, the indicator platform is also used to: receive the core indicator calculation caliber and the abnormal data threshold of the calculation caliber from user-defined business data; and associate the core indicator calculation caliber and the abnormal data threshold of the calculation caliber with the corresponding business model.

[0060] In this embodiment, it should be noted that users can manually define the calculation caliber of core business data metrics and the threshold for abnormal data in the metrics platform. For example, when sales fluctuations exceed ±20%, a red warning will be automatically displayed in the data. Responding to user customization, the metrics platform associates the defined core metric calculation caliber and the threshold for abnormal data with the corresponding business model, ensuring that the metrics under that business model simultaneously carry the calculation logic and the anomaly judgment criteria. Therefore, when metadata is subsequently synchronized to the BI platform, the BI platform can perform metric calculations based on this calculation caliber and perform anomaly detection and judgment based on this threshold.

[0061] In this embodiment of the application, the indicator platform is also used to: after extracting metadata from the business model, convert the metadata into a standardized metadata format that the BI platform can recognize and unify the semantic tags of the fields.

[0062] In this embodiment, it should be noted that, as a low-code metrics platform, after automatically extracting metadata from the business model, including entities, fields, relationships, metric definitions, and permission configurations, the platform further converts it into a standardized metadata format that the BI platform can recognize, and unifies the semantic tags of the fields. For example, the "CUST_ID" field is automatically labeled as the "customer identifier" semantic type, and the "SALE_INCOME" field is labeled as the "sales revenue" semantic type.

[0063] In one embodiment, a computer device is provided, which may be a server, and its internal structure diagram may be as follows: Figure 4As shown. The computer device includes a processor A01, a network interface A02, memory (not shown), and a database (not shown) connected via a system bus. The processor A01 provides computing and control capabilities. The memory includes internal memory A03 and a non-volatile storage medium A04. The non-volatile storage medium A04 stores an operating system B01, a computer program B02, and a database (not shown). The internal memory A03 provides an environment for the operation of the operating system B01 and the computer program B02 stored in the non-volatile storage medium A04. The database stores analytical data related to financial data. The network interface A02 is used for communication with external terminals via a network connection. When the computer program B02 is executed by the processor A01, it performs financial data analysis.

[0064] Those skilled in the art will understand that Figure 4 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0065] This application provides an apparatus including a processor, a memory, and a program stored in the memory and executable on the processor. When the processor executes the program, it performs the following steps: receiving business analysis text input by a user; parsing the business analysis text to obtain the business object, business time range, and business analysis type, and generating an analysis task; initiating an analysis request to a BI platform based on the analysis task, wherein the BI platform is configured with metadata and anomaly data models corresponding to the user's required business model; receiving a result dataset returned by the BI platform in response to the analysis request, wherein the result dataset includes business objects and business metrics, and the business metrics at least include business data, anomalous data, and the anomalous magnitude, dimension, and cause of the anomalous data; the result dataset is obtained by the BI platform after sequentially performing data query and anomalous data analysis on the analysis task; data query refers to the BI platform obtaining business data and performing data calculations based on metadata semantic mapping; anomalous data analysis refers to the BI platform performing anomaly detection on the business data obtained after executing the data query based on the anomalous data model; and generating and outputting analysis result text based on the result dataset.

[0066] This application also provides a computer program product, which, when executed on a data processing device, is suitable for executing an initialization program with the following method steps: receiving business analysis text input by a user; parsing the business analysis text to obtain the business object, business time range, and business analysis type, and generating an analysis task; initiating an analysis request to a BI platform based on the analysis task, wherein the BI platform is configured with metadata and anomaly data models corresponding to the business model required by the user; receiving a result dataset returned by the BI platform in response to the analysis request, wherein the result dataset includes business objects and business indicators, and the business indicators include at least business data, anomalous data, and the anomalous magnitude, dimension, and cause of the anomalous data; the result dataset is obtained by the BI platform after sequentially performing data query and anomalous data analysis on the analysis task, wherein data query refers to the BI platform obtaining business data and performing data calculations based on semantic mapping of metadata, and anomalous data analysis refers to the BI platform performing anomaly detection on the business data obtained after executing data query based on the anomalous data model; and generating and outputting analysis result text based on the result dataset.

[0067] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0068] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0069] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0070] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0071] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.

[0072] Memory may include non-persistent memory in computer-readable media, such as random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM. Memory is an example of computer-readable media.

[0073] Computer-readable media includes both permanent and non-permanent, removable and non-removable media that can store information using any method or technology. Information can be computer-readable instructions, data structures, modules of programs, or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, CD-ROM, digital versatile optical disc (DVD) or other optical storage, magnetic tape, magnetic disk storage or other magnetic storage devices, or any other non-transferable medium that can be used to store information accessible by a computing device. As defined herein, computer-readable media does not include transient computer-readable media, such as modulated data signals and carrier waves.

[0074] It should also be noted that the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element.

[0075] The above are merely embodiments of this application and are not intended to limit the scope of this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of the claims of this application.

Claims

1. A method for analyzing financial data, characterized in that, Applied to an intelligent agent platform, which interacts with a BI platform, the analysis method includes: Receive business analysis text input by the user; The business analysis text is parsed to obtain the business object, business time range, and business analysis type, and an analysis task is generated. Based on the analysis task, an analysis request is initiated to the BI platform, wherein the BI platform is configured with metadata and anomaly data models corresponding to the business model required by the user. The system receives a result dataset returned by the BI platform in response to the analysis request. This result dataset includes business objects and business metrics. The business metrics include at least business data, abnormal data, and the magnitude, dimension, and cause of the abnormal data. The result dataset is obtained by the BI platform after sequentially performing data queries and abnormal data analysis on the analysis task. Data query refers to the BI platform performing semantic mapping based on the metadata to obtain business data and then performing data calculations. Abnormal data analysis refers to the BI platform performing anomaly detection on the business data obtained after performing data queries based on the abnormal data model. The analysis results text is generated and output based on the resulting dataset.

2. The method for analyzing financial data according to claim 1, characterized in that, The process of parsing the business analysis text to obtain the business object, business time range, and business analysis type, and generating analysis tasks includes: Parse the business analysis text to obtain the business object, business time range and business analysis type, and determine whether the business analysis text is the user's first business analysis text. If the business analysis text is the initial business analysis text, an analysis task is generated; If the business analysis text is not the first business analysis text, retrieve the user's historical business analysis texts, identify the historical business objects in the historical business analysis texts as the business objects, and generate an analysis task.

3. The method for analyzing financial data according to claim 1, characterized in that, The intelligent agent platform is connected to the client, and the analysis method further includes: If abnormal data is found in the result dataset, an abnormal data alert is generated and pushed to the client.

4. A method for analyzing financial data, characterized in that, Applied to a BI platform, which interacts with an agent platform and a metrics platform, the analysis method includes: The system receives metadata sent by the metrics platform and constructs an abnormal data model based on the metadata. The metadata is extracted by the metrics platform from the business model after the user constructs the business model based on the data definition component provided by the metrics platform according to the required business scenario. In response to the received analysis request from the intelligent agent platform, the analysis task is semantically mapped based on the metadata to retrieve business data, and data calculations are performed on the business data. Based on the aforementioned abnormal data model, anomaly detection is performed on the calculated business data, and a result dataset for the anomaly-detected business data is generated. The result dataset includes business objects and business metrics. The business metrics include at least business data, abnormal data, and the magnitude, dimension, and cause of the anomaly in the abnormal data. The resulting dataset is returned to the agent platform.

5. The method for analyzing financial data according to claim 4, characterized in that, Based on the aforementioned anomaly data model, anomaly detection is performed on the calculated business data, and a result dataset for the anomaly-detected business data is generated, including: Based on the aforementioned abnormal data model, determine whether the calculated business data contains anomalies; If there are anomalies in the calculated business data, determine the dimension, magnitude, and cause of the anomaly in the database table, and generate a result dataset based on the anomaly and its magnitude, dimension, and cause. If there are no anomalies in the business data, a result dataset is generated based on normal business data.

6. The method for analyzing financial data according to claim 4, characterized in that, The metadata includes at least one of entities, fields, relationships, metric definitions, and permission configurations. The BI platform maps the metadata in the metric platform to metadata that conforms to the user's business behavior.

7. A financial data analysis system, characterized in that, The analysis system includes: An intelligent agent platform is used to receive business analysis text input by a user; parse the business analysis text to obtain the business object, business time range, and business analysis type, and generate an analysis task; initiate an analysis request to a BI platform based on the analysis task, wherein the BI platform is configured with metadata and anomaly data models corresponding to the business model required by the user; receive the result dataset returned by the BI platform in response to the analysis request, wherein the result dataset includes business objects and business indicators, and the business indicators include at least business data, anomaly data, and the anomaly magnitude, dimension, and cause of the anomaly data; and generate and output analysis result text based on the result dataset. The BI platform receives metadata sent by the metrics platform and constructs an anomaly data model based on the metadata; in response to an analysis request sent by the intelligent agent platform, it performs semantic mapping on the analysis task based on the metadata to retrieve business data and performs data calculations on the business data; it performs anomaly detection on the calculated business data based on the anomaly data model and generates a result dataset for the anomaly-detected business data; and it returns the result dataset to the intelligent agent platform.

8. The financial data analysis system according to claim 7, characterized in that, The analysis system also includes: The metrics platform provides data definition components to respond to users who construct business models based on the data definition components according to the required business scenarios, extract metadata from the business models, and send the metadata to the BI platform.

9. The financial data analysis system according to claim 8, characterized in that, The indicator platform is also used for: The core metrics calculation methods and abnormal data thresholds for receiving user-defined business data; The calculation methods for the core indicators and the thresholds for abnormal data in the calculation methods are associated with the corresponding business models.

10. The financial data analysis system according to claim 8, characterized in that, The indicator platform is also used for: After extracting metadata from the business model, the metadata is converted into a standardized metadata format that the BI platform can recognize and the semantic tags of the fields are unified.

Images