Intelligent service analysis optimization method and system based on multi-dimensional data warehouse

Abstract

The invention belongs to the field of data processing, and particularly relates to an intelligent service analysis optimization method and system based on a multi-dimensional data warehouse, and the method comprises the steps: obtaining the data attribute information of a to-be-recognized object for the to-be-processed object loaded in the data warehouse; inputting the data attribute information into a dynamic data analysis model, and performing data identity recognition on the to-be-recognized object to obtain target identity information of the to-be-recognized object; based on the target identity information, performing identity tracing on the to-be-identified object by adopting a dynamic tracing model to obtain target tracing information of the to-be-identified object; and based on the target identity information and the target traceability information, triggering a target data service for the to-be-identified object to complete intelligent service analysis of the to-be-processed object. According to the method, the data identity recognition and traceability capability can be improved, powerful support is provided for intelligent service analysis, and the method is dedicated to realizing data-driven decision making.

Description

Technical Field

[0001] The present application relates to the field of data processing, and in particular to an intelligent business analysis optimization method and system based on a multidimensional data warehouse. Background Art

[0002] A data warehouse is a system used to store, manage, and analyze large amounts of data, designed to support the business intelligence (BI) capabilities of an enterprise. The data structure of a data warehouse is built around specific topics, such as sales, finance, customers, etc. This structure allows users to easily perform analysis in specific areas without having to focus on specific applications. The data warehouse integrates data from multiple data sources, allowing enterprises to efficiently perform data analysis, reporting, and decision support. Data comes from multiple different sources, and after cleaning and conversion, it is uniformly stored in the data warehouse. This allows users to get a consistent view of the data without having to worry about where the data comes from.

[0003] In the related art, data warehouses usually need to integrate data from multiple sources. The process of tracking data flows and changes may involve multiple conversions and merges, which increases the complexity of the overall system and makes tracking and debugging difficult, especially when encountering data inconsistencies or errors. Since many traditional data warehouses are usually updated in batches and on a timed basis, this may lead to insufficient real-time data tracking. Users may not be able to quickly obtain the latest data change status, affecting the timeliness of decision-making, especially in a rapidly changing business environment, where real-time data is critical. In addition, different data sources may use different standards and formats for data tracking, resulting in difficulties in information integration. In the process of data tracking, data redundancy may occur, especially when integrating data from different sources. If multiple users or applications are using the same data set for analysis, information inconsistencies may occur, increasing management complexity.

[0004] Therefore, in order to solve at least one of the above technical problems, it is urgent to propose a new intelligent business analysis optimization solution based on a multidimensional data warehouse. Summary of the invention

[0005] This application provides an intelligent business analysis optimization method and system based on a multidimensional data warehouse to improve data identity recognition and traceability capabilities, and also provides strong support for intelligent business analysis, and is committed to achieving data-driven decision-making.

[0006] In a first aspect, the present application provides an intelligent business analysis optimization method based on a multidimensional data warehouse, the method comprising:

[0007] For the object to be processed loaded in the data warehouse, data attribute information of the object to be identified is obtained; the object to be identified is data information from at least one data source in the multidimensional data source; the data attribute information at least includes: data content, data type, data source identifier, data source type, and data collection method of the object to be identified;

[0008] Inputting the data attribute information into a dynamic data analysis model, performing data identity recognition on the object to be identified, and obtaining target identity information of the object to be identified;

[0009] Based on the target identity information, a dynamic tracing model is used to trace the identity of the object to be identified, so as to obtain the target tracing information of the object to be identified;

[0010] Based on the target identity information and the target traceability information, a target data service for the object to be identified is triggered to complete intelligent service analysis of the object to be processed.

[0011] In a second aspect, an embodiment of the present application provides an intelligent business analysis and optimization device based on a multidimensional data warehouse, the device comprising:

[0012] The acquisition unit is configured to acquire data attribute information of the object to be identified for the object to be processed loaded in the data warehouse; the object to be identified is data information from at least one data source in the multidimensional data source; the data attribute information at least includes: data content, data type, data source identifier, data source type, and data collection method of the object to be identified;

[0013] an identification unit configured to input the data attribute information into a dynamic data analysis model, perform data identity identification on the object to be identified, and obtain target identity information of the object to be identified;

[0014] A tracing unit is configured to perform identity tracing on the object to be identified based on the target identity information by using a dynamic tracing model to obtain target tracing information of the object to be identified;

[0015] The trigger unit is configured to trigger the target data service for the object to be identified based on the target identity information and the target traceability information, so as to complete the intelligent service analysis of the object to be processed.

[0016] In a third aspect, an embodiment of the present application provides a computing device, the computing device comprising:

[0017] at least one processor, memory, and input-output unit;

[0018] The memory is used to store computer programs, and the processor is used to call the computer programs stored in the memory to execute the intelligent business analysis and optimization method based on a multidimensional data warehouse of the first aspect.

[0019] In a fourth aspect, a computer-readable storage medium is provided, which includes instructions, and when the instructions are executed on a computer, the computer executes the intelligent business analysis and optimization method based on a multidimensional data warehouse of the first aspect.

[0020] In the technical solution provided by the embodiment of the present application, first, for the objects to be processed loaded in the data warehouse, the data attribute information of the objects to be identified is obtained. Among them, the objects to be identified are data information from at least one data source in the multidimensional data source. The data attribute information includes at least: the data content, data type, data source identifier, data source type, and data collection method of the object to be identified. Then, the data attribute information is input into the dynamic data analysis model, and the data identity of the object to be identified is performed to obtain the target identity information of the object to be identified. Next, based on the target identity information, a dynamic traceability model is used to trace the identity of the object to be identified to obtain the target traceability information of the object to be identified. Finally, based on the target identity information and the target traceability information, the target data business for the object to be identified is triggered to complete the intelligent business analysis of the object to be processed.

[0021] In the technical solution of this application, the implementation of dynamic data analysis and identity traceability processes in a multidimensional data warehouse can bring many beneficial effects. First, by comprehensively integrating the data attribute information of the object to be identified and inputting it into the dynamic data analysis model, the rich information in the multidimensional data source can be used to significantly improve the accuracy of identity recognition. The dynamic analysis model adapts to the ever-changing data characteristics and reduces the probability of misidentification and missed identification. Secondly, the use of a dynamic traceability model can effectively track the source and changes of data, clarify the reliability and change history of data, thereby improving data traceability capabilities and providing support for data supervision and compliance review. In addition, combining target identity information with target traceability information can promote intelligent business analysis, provide a basis for decision support strategies, such as personalized recommendations and risk assessment, and improve business operation efficiency. Dynamic traceability management can also enhance data security and transparency, ensure that the source of data is clear and traceable, enhance users' trust in data management, and promptly discover potential security risks and respond quickly. By integrating multidimensional data sources, the correlation and potential value between data can be mined to help companies conduct market analysis and consumer behavior analysis, so as to better meet user needs and optimize products and services. In addition, the dynamic data analysis and traceability model enables enterprises to adjust analysis and decision-making strategies in real time according to market changes, quickly respond to changes and seize market opportunities, and enhance competitive advantages. Finally, through automated data identification and traceability, human intervention is reduced, data processing efficiency is improved, and collaboration between departments is promoted. In summary, the dynamic data analysis optimization method based on the multidimensional data warehouse not only improves the data identification and traceability capabilities, but also provides strong support for intelligent business analysis, and is committed to achieving data-driven decision-making and promoting sustainable development and growth of the business. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] The drawings described herein are used to provide a further understanding of the present application and constitute a part of the present application. The illustrative embodiments of the present application and their descriptions are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings:

[0023] Figure 1 It is a flowchart of an intelligent business analysis optimization method based on a multidimensional data warehouse according to an embodiment of the present application;

[0024] Figure 2 It is a schematic diagram of the principle of an intelligent business analysis and optimization method based on a multidimensional data warehouse according to an embodiment of the present application;

[0025] Figure 3 It is a structural diagram of an intelligent business analysis and optimization device based on a multidimensional data warehouse according to an embodiment of the present application;

[0026] Figure 4 It is a structural schematic diagram of an electronic device according to an embodiment of the present application. DETAILED DESCRIPTION

[0027] In order to make the purpose, technical solution and advantages of the embodiments of the present application clearer, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0028] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those commonly understood by those skilled in the art to which this application belongs. The terms used herein in the specification of this application are only for the purpose of describing specific embodiments and are not intended to limit this application.

[0029] A data warehouse is a system used to store, manage, and analyze large amounts of data, designed to support the business intelligence (BI) capabilities of an enterprise. The data structure of a data warehouse is built around specific topics, such as sales, finance, customers, etc. This structure allows users to easily perform analysis in specific areas without having to focus on specific applications. The data warehouse integrates data from multiple data sources, allowing enterprises to efficiently perform data analysis, reporting, and decision support. Data comes from multiple different sources, and after cleaning and conversion, it is uniformly stored in the data warehouse. This allows users to get a consistent view of the data without having to worry about where the data comes from.

[0030] In the related art, data warehouses usually need to integrate data from multiple sources. The process of tracking data flows and changes may involve multiple conversions and merges, which increases the complexity of the overall system and makes tracking and debugging difficult, especially when encountering data inconsistencies or errors. Since many traditional data warehouses are usually updated in batches and on a timed basis, this may lead to insufficient real-time data tracking. Users may not be able to quickly obtain the latest data change status, affecting the timeliness of decision-making, especially in a rapidly changing business environment, where real-time data is critical. In addition, different data sources may use different standards and formats for data tracking, resulting in difficulties in information integration. In the process of data tracking, data redundancy may occur, especially when integrating data from different sources. If multiple users or applications are using the same data set for analysis, information inconsistencies may occur, increasing management complexity.

[0031] To solve at least one of the above technical problems, an embodiment of the present application provides an intelligent business analysis optimization method and system based on a multidimensional data warehouse.

[0032] Specifically, first, for the objects to be processed loaded in the data warehouse, the data attribute information of the objects to be identified is obtained. Among them, the objects to be identified are data information from at least one data source in the multidimensional data source. The data attribute information includes at least: the data content, data type, data source identifier, data source type, and data collection method of the object to be identified. Then, the data attribute information is input into the dynamic data analysis model, and the data identity of the object to be identified is performed to obtain the target identity information of the object to be identified. Next, based on the target identity information, the dynamic traceability model is used to trace the identity of the object to be identified to obtain the target traceability information of the object to be identified. Finally, based on the target identity information and the target traceability information, the target data business for the object to be identified is triggered to complete the intelligent business analysis of the object to be processed.

[0033] In the technical solution of this application, intelligent business analysis based on a multidimensional data warehouse aims to improve the information management and analysis efficiency of the objects to be processed in the data warehouse, especially to solve the problems of complexity, lack of real-time performance, data redundancy and consistency, lack of standardization and security risks mentioned above. First, rich attribute information of the objects to be processed is extracted from the multidimensional data source, including data content, data type, data source identifier, data source type and data collection method. The data is structured and normalized through standardized data attribute information acquisition. This reduces the possibility of data redundancy and improves data consistency. In addition, detailed records of attribute information also help to track data in subsequent analysis, and data from different sources are clear at a glance. Then, the acquired data attribute information is input into a dynamic data analysis model, which performs data identity recognition for the objects to be identified. Using dynamic analysis, the model can be continuously adjusted to adapt to new data features and changes, and improve real-time performance. The dynamic analysis model can identify changes in data and provide immediate feedback, thereby quickly responding to changes in market or business needs and avoiding delays. Then, based on the identity information to be verified, a dynamic traceability model is used to trace the identity of the object to be identified and obtain the target traceability information. Through the implementation of the dynamic traceability model, the source and change trajectory of the data can be traced to ensure the integrity and source transparency of the data. This enhances the security of the data, can identify potential data leakage risks, and issue early warnings for problems, reducing the risk of non-compliance. Furthermore, based on the target identity information and target traceability information, the data business for the object to be identified can be automatically triggered to complete intelligent business analysis. This step closely combines data analysis with business logic to achieve automation and intelligence of the business. This reduces manual intervention in the analysis process, reduces complexity, and improves response speed, making intelligent business analysis more efficient and reliable. At the same time, the automatic triggering mechanism ensures the synchronization between data updates and analysis, avoiding outdated or inconsistent data analysis results.

[0034] Through the technical solution of this application, intelligent business analysis based on dynamic analysis models and traceability models is realized. By standardizing data attribute information and dynamic analysis, data redundancy is reduced and data consistency is ensured. The application of dynamic models allows real-time processing and analysis of data to quickly respond to business changes. By tracking and managing identity and traceability information, the transparency of data use is improved and the risk of data leakage is reduced. The automated business analysis process reduces dependence on manual intervention, making the management and analysis process more concise and efficient. Thereby, an effective optimization method is provided for intelligent business analysis in multidimensional data warehouses, helping enterprises better respond to business needs and market competition.

[0035] The intelligent business analysis optimization scheme based on a multidimensional data warehouse provided in the embodiment of the present application can be executed by an electronic device, which can be a server, a server cluster, or a cloud server. The electronic device can also be a terminal device such as a mobile phone, a computer, a tablet computer, a wearable device, or a dedicated device (such as a dedicated terminal device with an intelligent business analysis optimization system based on a multidimensional data warehouse, etc.). In an optional embodiment, a service program for executing an intelligent business analysis optimization scheme based on a multidimensional data warehouse can be installed on the electronic device.

[0036] Figure 1 A schematic diagram of an intelligent business analysis optimization method based on a multidimensional data warehouse provided in an embodiment of the present application is shown in FIG. Figure 1 As shown, the method comprises the following steps:

[0037] 101. For an object to be processed loaded in a data warehouse, obtain data attribute information of the object to be identified;

[0038] 102, inputting the data attribute information into a dynamic data analysis model, performing data identity recognition on the object to be recognized, and obtaining target identity information of the object to be recognized;

[0039] 103, based on the target identity information, using a dynamic tracing model to trace the identity of the object to be identified, to obtain target tracing information of the object to be identified;

[0040] 104. Based on the target identity information and the target traceability information, trigger a target data service for the object to be identified to complete intelligent service analysis of the object to be processed.

[0041] The multidimensional data warehouse dynamic data analysis and identity tracing process in the embodiment of the present application is a complete data processing process that integrates data attribute acquisition, data analysis, identity recognition, tracing analysis and business triggering. First, the objects to be processed loaded in the data warehouse need to obtain their detailed data attribute information, which includes but is not limited to data content (such as user comment text, purchase amount, uploaded pictures, sensor readings), data type (such as descriptive text type, numerical type, picture type, time type), data source identifier (such as user database, e-commerce platform API, sensor device), data source type (such as relational database, non-relational database, API interface, file system) and data collection method (such as automatic collection, manual input, batch import, real-time processing).

[0042] Next, the acquired data attribute information is input into the dynamic data analysis model, which performs data identity recognition on the object to be recognized through steps such as feature extraction (such as keywords in comments and statistical characteristics of sales), model training and optimization (using historical data to train machine learning models and adjusting model parameters through optimization algorithms) and identity recognition (using the trained model to recognize the object to be recognized and obtain target identity information, such as user identity and product category).

[0043] After obtaining the target identity information, a dynamic traceability model is used to trace the identity of the object to be identified. This model achieves identity traceability through steps such as data source tracking (tracking the historical source of data and identifying changes in data in different data sources), data change records (recording the historical status of data and its change records, clarifying the reliability and change history of data), and traceability analysis (obtaining the target traceability information of the object to be identified, such as the original source of the data and the modification history of the data, by analyzing the historical status and change records of the data).

[0044] Based on the target identity information and target traceability information, the target data business for the object to be identified is triggered to complete intelligent business analysis. The specific steps include business rule matching (matching the corresponding business rules according to the target identity information and target traceability information, such as risk control and personalized recommendation), intelligent decision support (formulating intelligent decision support strategies according to the matched business rules, such as recommending related products based on the user's historical purchase records and assessing the user's credit risk) and business execution (executing the corresponding business operations, such as generating personalized recommendation lists and generating risk assessment reports).

[0045] In the embodiment of the present application, the object to be identified is data information from at least one data source in the multidimensional data source. The integrated processing of multidimensional data sources can dig out the correlation and potential value between data and provide support for business. For example, the dynamic analysis model can reveal trends and patterns between different data dimensions, helping business parties to conduct market analysis, consumer behavior analysis, etc., so as to better meet user needs and optimize products and services.

[0046] Further optionally, the data attribute information includes at least: data content, data type, data source identifier, data source type, and data collection method of the object to be identified.

[0047] In the process of implementing dynamic data analysis and identity traceability of multidimensional data warehouse, the data attribute information of the object to be identified can include multiple aspects. First, the data content refers to the specific information or value of the object to be identified, such as the user's message "I am very satisfied with this product!", the user's purchase amount "99.99 yuan", the uploaded avatar picture or product photo, and the temperature sensor reading "25.5℃". Secondly, the data type helps the system understand the data format, which may include text type (such as user comments), numerical type (such as sales "10000"), image type (such as image files in JPEG or PNG format) and time type (such as timestamp "2023-09-29 12:34:56"). In terms of data source identification, a unique identifier can be used to distinguish the data source, such as "user_db_001" for the user database, "ecommerce_api_002" for the e-commerce platform API, and "sensor_12345" as the unique identifier of the sensor device. In addition, the data source type describes the way data is stored or obtained, which may be a relational database (such as MySQL or PostgreSQL), a non-relational database (such as MongoDB), an API interface (such as RESTful API) or a file system (such as CSV or Excel file). Finally, the data collection method describes the process of data collection or generation, such as automatically collecting data from sensors through a program, users manually entering comments on a web page, batch importing historical sales data, or real-time stream processing through a data stream platform. In summary, assuming that the object to be identified is a user comment on an e-commerce platform, its specific data attribute information can be presented as follows: the data content is "This phone is so easy to use!", the data type is text, the data source identifier is "ecommerce_db_001", the data source type is a relational database (MySQL), and the data collection method is manual input. With this information, the system can effectively identify and trace the object to be identified, providing support for subsequent business analysis and decision-making.

[0048] Exemplarily, in 101, for the object to be processed loaded in the data warehouse, data attribute information of the object to be identified is obtained.

[0049] In the data warehouse, for the loaded objects to be processed, the process of obtaining their data attribute information can be concretized into the following steps:

[0050] For example, for a user review text, its data content is: "This phone is so useful!" For other objects, such as the user's purchase amount of "99.99 yuan", the user's uploaded avatar is a JPEG image file, and the temperature sensor reading is "25.5℃". The data type of the review text is descriptive text type ("text"); the user's purchase amount type is numeric type ("decimal"); the avatar image data type is image type ("image / jpeg"); the temperature sensor reading data type is numeric type ("float"). The above review text comes from the user database, identified as "user_db_001"; the purchase amount comes from the e-commerce API, identified as "ecommerce_api_002"; the avatar image also comes from "user_db_001"; and the temperature reading comes from the sensor identified as "sensor_12345". The data source type of the comment text is a relational database (MySQL); the data source type of the purchase amount is an API interface (RESTful API); the data source type of the avatar image is a relational database (PostgreSQL); and the data source type of the temperature reading is an API interface (IoT sensor API). The comment text is obtained by users manually entering it on the e-commerce platform; the purchase amount is obtained through automatic collection by the system; the avatar image is manually uploaded by the user; and the temperature reading is obtained through automatic collection.

[0051] Through these steps, it is clearly demonstrated how to obtain the data attribute information of the object to be processed in the data warehouse. This information provides the basis for subsequent data analysis, identity identification, traceability analysis and business triggering.

[0052] As an optional embodiment, it is assumed that the dynamic data analysis model includes at least: a first feature extraction layer, an analysis layer, a prediction layer, and an optimization layer.

[0053] In a dynamic data analysis model, the design of each layer is crucial to the data identity recognition process. First, the main function of the first feature extraction layer is to extract features that can be used for subsequent analysis from the input raw data attribute information. This layer includes data individual attribute feature extraction, such as user behavior, purchase records, social network interactions, geographic location data, etc.; feature encoding, converting the extracted features into a format suitable for model processing; and data preprocessing, such as missing value filling, outlier detection and normalization processing to ensure data quality and consistency. Next is the analysis layer. After feature extraction, the data will be further analyzed. The main functions include key feature analysis, using statistical and machine learning methods to identify the features most relevant to identity recognition. This layer uses feature selection algorithms, such as LASSO regression and random forest, to screen out the attributes that have the greatest impact on identity feature prediction, and assign importance weights to each key feature to support subsequent more accurate identity feature prediction. The prediction layer is responsible for predicting identity features based on the key attribute features obtained by the analysis layer, applying machine learning algorithms such as logistic regression, support vector machine (SVM), deep learning, etc. to model key attribute features, calculate the probability distribution of possible identity features of the object to be identified, and generate multiple candidate identities and their corresponding confidence levels. Finally, there is the optimization layer, whose responsibility is to process and optimize the prediction results, including using non-maximum suppression to remove identity information with lower probability values ​​and retain the target identity information with the highest probability, thereby significantly reducing the error rate and noise in identity recognition. In addition, this layer will continue to adjust and optimize model parameters and structures based on feedback from actual applications to improve the accuracy and robustness of identity recognition. Through the collaborative work of the first feature extraction layer, analysis layer, prediction layer, and optimization layer, the dynamic data analysis model can efficiently identify the target identity information of the object to be identified from the data attribute information. This structured hierarchical design not only improves the accuracy of identity recognition, but also provides a good foundation for subsequent expansion and optimization.

[0054] Based on the above assumptions, in step 102, the data attribute information is input into the dynamic data analysis model, data identity recognition is performed on the object to be identified, and the target identity information of the object to be identified is obtained, which can be implemented as follows:

[0055] Extracting, through the first feature extraction layer, data individualized attribute features corresponding to the object to be identified from the data attribute information;

[0056] By means of the analysis layer, key feature analysis is performed on the individualized attribute features of the data to obtain key attribute features corresponding to the individualized attribute features of the data;

[0057] Through the prediction layer, identity feature prediction is performed according to the key attribute features to obtain a probability distribution of candidate identities of the object to be identified;

[0058] Through the optimization layer, non-maximum suppression processing is adopted to select the target identity information with the highest identity probability value.

[0059] In the dynamic data analysis model, in order to achieve data identification of the object to be identified and obtain its target identity information, the whole process involves four main levels: the first feature extraction layer, the analysis layer, the prediction layer, and the optimization layer. First, in the first feature extraction layer, the goal is to extract the individual attribute characteristics of the object to be identified from the input data attribute information, including converting user behavior, social interaction, purchase records and other information into usable features, such as sentiment scores or activity. At the same time, natural language processing technology is used to extract keywords and sentiment polarity from text data, and data preprocessing is performed to improve data consistency and availability.

[0060] Next is the analysis layer, which is responsible for in-depth analysis of the extracted individual attribute features. Specifically, it uses statistical analysis and machine learning methods to identify important features related to the target identity, uses feature selection algorithms (such as LASSO regression and random forest) to determine the features that have the greatest impact on identity recognition, and assigns them importance scores so that subsequent models can focus on these key features.

[0061] At the prediction layer, the model will predict identity features based on the key attribute features obtained at the analysis layer, and will be trained by selecting appropriate machine learning algorithms (such as logistic regression, support vector machine, decision tree or deep learning). After training, the model outputs the identity probability distribution of the object to be identified based on the key features, generates multiple candidate identities and their confidence levels, and reflects the degree of certainty for each identity.

[0062] Finally, in the optimization layer, the prediction results are processed and optimized, especially the selection of candidate identities. By applying non-maximum suppression, the identities with the highest recognition probability are screened out, reducing misjudgments and noise, thereby improving the accuracy of identity recognition. In addition, based on feedback from actual applications, the model will continuously adjust parameters and structures to improve accuracy and robustness, and adapt to changing user behaviors through continuous optimization.

[0063] Through the collaborative work of these four levels, the dynamic data analysis model can efficiently extract and identify the target identity information of the object to be identified from the data attribute information, thereby improving the accuracy of identity recognition and providing strong support for applications in different scenarios.

[0064] Further optionally, the probability distribution of the candidate identities of the object to be identified is expressed as follows:

[0065]

[0066] Among them, P(I jx) represents the probability of the candidate identity of the jth object to be identified under the key attribute feature x. j , w k represents the first weight item corresponding to the key attribute feature x, b k represents the first bias item corresponding to the key attribute feature x, w j represents the second weight item corresponding to the key attribute feature x, b j Represents the second bias term corresponding to the key attribute feature x.

[0067] Through the above calculation examples, we can see how to apply specific weights and biases to the data of the object to be identified and calculate the probability distribution of its identity features. By comprehensively considering the importance and interaction effects of each feature, a more accurate identity recognition result can be obtained.

[0068] Further optionally, in 102, after the data attribute information is input into the dynamic data analysis model, data identity recognition is performed on the object to be identified, and the target identity information of the object to be identified is obtained, it is also possible to verify whether the target identity information is consistent with the reference identity information indicated by the preset verification information contained in the data attribute information. Among them, the preset verification information is at least one of the data type, data source identifier, data source type, and data collection method. The reference identity information is an identity type that is pre-set in the preset identity verification library and bound to the preset verification information. Furthermore, if the target identity information is consistent with the reference identity information, it is determined that the target identity information has been verified. If the target identity information is inconsistent with the reference identity information, it triggers the step of re-executing the step of inputting the data attribute information into the dynamic data analysis model, performing data identity recognition on the object to be identified, and obtaining the target identity information of the object to be identified.

[0069] For example, on an e-commerce platform, the identity of a customer is identified through user behavior data, and the consistency between the identification result and the preset identity information needs to be verified. First, the user behavior data of the object to be identified includes purchase records and browsing history, the data type is "behavior data", the data source identifier is "user ID_123", the data source type is "e-commerce platform", and the data collection method is "API". Then, the dynamic data analysis model analyzes these behavior data and identifies the target identity information as "VIP user". In this process, the preset verification information includes data type, data source identifier, data source type and data collection method, and is bound to the reference identity information, such as the identity type is "VIP user". Subsequently, the system will verify the consistency of the target identity information and the reference identity information. If the two are consistent, the identity authentication is passed, the system records the user identity and grants the corresponding VIP rights; on the contrary, if the target identity identified by the model is "ordinary user" and the reference information is still "VIP user", it will trigger the step of re-execution of identity identification. At this time, the system will re-enter the data attribute information and use the dynamic data analysis model to analyze the user behavior data again, in order to obtain more accurate identity information, while ensuring data quality to improve the recognition accuracy. Through this process, the platform can ensure the reliability and accuracy of identity identification, effectively prevent incorrect authentication, and protect user interests.

[0070] Further optionally, in 102, the data attribute information is input into the dynamic data analysis model, and the data identity of the object to be identified is identified. After obtaining the target identity information of the object to be identified, it is also possible to verify whether the object to be identified has the storage permission of the target storage area. If the object to be identified does not have the storage permission, the dynamic traceability model is used to trace the identity of the object to be identified based on the target identity information to obtain the target traceability information of the object to be identified.

[0071] Wherein, the verification of the storage permission includes at least data attributes, access level, time range, source category; the source category includes at least sensitive data and non-sensitive data. Furthermore, the time range includes at least: data valid range, data invalid range, data custom effective range. In the data management system, verifying whether the object to be identified has storage permission for the target storage area is a multi-dimensional process. Data attributes refer to the characteristics or types of data. In storage permission verification, data attributes are the first aspect to be checked. For example, example attributes, user behavior data, financial data, customer information, etc. The verification function ensures that users can only access data types that are consistent with their permissions.

[0072] Access levels define the user's access rights to data, usually including read, write, delete and other operations. Access level verification ensures that users can only perform operations within their authority. For example, sample access levels, read permission (only allows viewing data), write permission (allows modification and addition of data), delete permission (allows deletion of data), verification role, to prevent unauthorized operations, such as ordinary users trying to delete sensitive data.

[0073] The time range defines the effective time period of the user's access rights to the data. Time range verification ensures that users can only access data within the specified time, including the following types: Data valid range, which specifies the time period during which users can access data, such as "2023-01-01" to "2024-12-31". Data invalid range, which specifies the time period during which users are not allowed to access data, such as after "2025-01-01". Data custom effective range, which is a custom time period set according to specific needs, such as "2024-01-01" to "2024-12-31". The verification function controls the timeliness of data access and prevents expired access or data operations during non-working hours.

[0074] Source categories refer to specific categories of data, which are usually divided into sensitive data and non-sensitive data. Source category verification ensures that users can only access data categories that match their permissions. For example, sensitive data includes user personal information, financial records, medical information, etc. Non-sensitive data includes browsing history, purchase records, log information, etc. Verification prevents sensitive data from being accessed by unauthorized users and enhances data security.

[0075] Suppose there is a data management system. In a data management system, the specific steps for storage permission verification of the object to be identified include four aspects. The first is data attribute verification, which ensures that the object to be identified can only access "user behavior data" but cannot access financial data or customer information. Secondly, access level verification is performed to confirm that the object to be identified can only perform "read" and "write" operations and cannot be deleted. Next is time range verification. By checking the valid range of the data (2023-01-01 to 2024-12-31), invalid range (after 2025-01-01) and custom effective range (2024-01-01 to 2024-12-31), ensure that the object to be identified only accesses the data within the valid time period and avoids operations within the invalid time period. Finally, source category verification is performed to ensure that the object to be identified can only access non-sensitive "browsing history" and cannot access sensitive "user personal information". Through comprehensive verification of data attributes, access levels, time ranges, and source categories, the data management system can effectively ensure that the identified objects can only access data consistent with their permissions, thus playing an important role in protecting sensitive data and maintaining system security.

[0076] In a data management system, after the data identity of the object to be identified is identified, it is verified whether it has the storage permission to access certain sensitive data. First, the data attribute information of the object to be identified includes the data attribute "user behavior data", the data source identifier "user ID_456", the data source type "e-commerce platform" and the data collection method "API". After the dynamic data analysis model analyzes this information, it identifies the target identity information as "administrator". Then, the system verifies the storage permission of the identity. The verification conditions include: data attribute "user behavior data", access level "read, write", time range for data validity range "2023-01-01" to "2024-12-31", data custom effective range "2024-01-01" to "2024-12-31" and source category (sensitive data "user personal information", non-sensitive data "browsing history"). The current time "2024-09-29" is within the valid range, and the administrator identity includes read and write permissions for sensitive data "user personal information" in the preset permission rules. If all conditions are matched, the identity has the permission to access all data in the target storage area. If the verification is successful, the system allows access; if the verification fails, such as the administrator does not have the permission to write certain sensitive data, the system will reject and record a notification. After the permission verification is passed, the system uses a dynamic traceability model based on the administrator identity to obtain the target traceability information (such as IP address, access time), and verifies its consistency with expectations. Through this mechanism, sensitive data security and permission management are ensured.

[0077] As an optional embodiment, it is assumed that the dynamic traceability model at least includes: a second feature extraction layer, a path construction layer, a mapping layer, an adjustment layer, and a traceability layer. Specifically, the dynamic traceability model is a multi-level structure, including a second feature extraction layer, a path construction layer, a mapping layer, an adjustment layer, and a traceability layer. Each layer has its own unique role and detailed process. First, the second feature extraction layer extracts the data flow characteristics of the object to be identified through the data flow mode corresponding to the target identity information, including data path characteristics, data conversion characteristics, and data operation behavior characteristics, and records the flow path, conversion operation, and specific behavior of the data in the system. Then, the path construction layer sets the path based on these characteristics to generate a first data flow path to reflect the specific flow mode of the data. Then, the mapping layer splits the data flow characteristics into flow path feature points and maps them to the spherical feature space to obtain flow high-dimensional features, which better represent the geometric relationship and similarity between feature points. The adjustment layer measures the similarity by calculating the cosine distance between the flow high-dimensional features, optimizes and corrects the features, and improves the accuracy and consistency of the features. Finally, the traceability layer inputs the optimized high-dimensional flow features into the relationship classifier, predicts the relationship classification distribution between the flow path feature points, and generates the target traceability information of the object to be identified based on the predicted distribution value, including the data flow path, conversion process and operation behavior. Through the collaborative work of these levels, the dynamic traceability model can comprehensively analyze and process the data flow characteristics of the object to be identified, generate target traceability information containing complex relationships, and provide strong support for data management and security.

[0078] Based on the above assumptions, in 103, based on the target identity information, a dynamic tracing model is used to trace the identity of the object to be identified, and the target tracing information of the object to be identified is obtained, which can be implemented as follows:

[0079] Through the feature extraction layer, according to the data flow mode corresponding to the target identity information, the feature extraction of the object to be identified is performed to obtain the data flow feature of the object to be identified; the data flow feature at least includes: data path feature, data conversion feature, and data operation behavior feature;

[0080] By means of the path construction layer, a path is set based on the data flow characteristics of the object to be identified, so as to obtain a first data flow path of the object to be identified;

[0081] Through the mapping layer, the data flow features of the object to be identified are split into flow path feature points, and the flow path feature points are mapped to the spherical feature space to obtain the flow high-dimensional features corresponding to the flow path feature points;

[0082] Obtaining the cosine distance measurement similarity of the flow high-dimensional features through the adjustment layer, and optimizing and correcting the flow high-dimensional features by using the cosine distance measurement similarity to obtain optimized flow high-dimensional features;

[0083] Through the traceability layer, the optimized high-dimensional flow features of the flow path feature points are input into the relationship classifier to obtain the relationship classification prediction distribution value between the flow path feature points, and the target traceability information of the object to be identified is generated based on the relationship classification prediction distribution value.

[0084] The dynamic traceability model comprehensively traces the identity of the object to be identified through its five levels of functions and generates detailed target traceability information. The feature extraction layer extracts the data flow characteristics of the object to be identified based on the target identity information and its corresponding data flow mode. The extracted features include data path features, data conversion features, and data operation behavior features. Specifically, the data path feature records the flow path of data in the system, reflecting the transmission process of data from the source to the target; the data conversion feature records the various conversion operations that occur during the flow process, such as format conversion and data cleaning; the data operation behavior feature records specific data operation behaviors, such as reading, writing, and deleting.

[0085] The path construction layer is based on the data flow features extracted from the object to be identified. This layer sets the path and generates the first data flow path. The generated path should include the starting point, intermediate nodes and target points of the data, as well as the data conversion and operation behaviors at each node. This step clarifies the specific flow mode and process of data in the system.

[0086] The mapping layer refines the data flow features of the object to be identified into multiple flow path feature points, and maps these feature points into the spherical feature space. In this way, high-dimensional flow features can be obtained, which can better represent the geometric relationship and similarity between feature points. The splitting and mapping of feature points is an important basis for subsequent analysis.

[0087] The adjustment layer is responsible for obtaining the cosine distance measurement similarity between the high-dimensional features of the flow, and optimizing and correcting the high-dimensional features of the flow. By calculating the cosine similarity between feature points, the similarity between them can be quantified, and adjustments can be made based on this similarity, thereby improving the accuracy and consistency of the features.

[0088] The traceability layer inputs the optimized high-dimensional flow features into the relationship classifier to predict the relationship classification distribution between the flow path feature points. Through the output of the relationship classifier, the probability distribution of the relationship between the feature points can be obtained, and based on this distribution, the final target traceability information of the object to be identified is generated, including the flow path, conversion process and operation behavior of the data.

[0089] In this way, through the collaborative work of these five levels, the dynamic traceability model can systematically analyze and process the data flow characteristics of the object to be identified, generate detailed target traceability information, reflect the complex relationship between data, and provide strong support for data management and security. This process covers a comprehensive process from feature extraction, path construction, mapping analysis, similarity optimization to relationship classification, ensuring the accuracy and practicality of traceability information.

[0090] Further optionally, the cosine distance metric of the flow high-dimensional feature is expressed as the following expression:

[0091] T i =S(i)(c·||ω i ||-α)cos(m·β i +d)

[0092] Among them, T i represents the cosine distance measurement similarity of the i-th flow high-dimensional feature vector, S(i) represents the i-th flow high-dimensional feature vector, α represents the cosine value amplification factor, c is the cosine value reduction factor, d is the cosine value balance adjustment factor, m is an adjustable hyperparameter, ω i represents the weight vector of the high-dimensional feature vector of the i-th flow, β i Represents the angle between the i-th flow high-dimensional feature vector and the weight vector.

[0093] In this way, the construction of the cosine distance metric for high-dimensional features plays an important role in the dynamic traceability model. In the cosine distance metric for high-dimensional features, each component complements each other, making the similarity calculation between feature vectors more accurate, flexible and adaptable. This efficient similarity evaluation mechanism has a key impact on identifying the objects to be processed and accurately generating target traceability information. By reasonably adjusting these parameters, the model can maintain superior performance in different application scenarios.

[0094] In 104, based on the target identity information and the target traceability information, a target data service for the object to be identified is triggered to complete intelligent service analysis of the object to be processed.

[0095] For example, the identity information of each participating account is automatically recorded during the transaction process, such as the basic information of the account holder, the historical transaction records of the account, etc. The transaction path and related operations of the account involved are tracked through the dynamic traceability model, and the complete transaction flow information is collected, including data path, conversion records and related operation behaviors.

[0096] In 104, according to business needs, define when to trigger the target data business. For example, when the transaction amount exceeds a certain threshold, or when an abnormal transaction pattern between accounts is identified. Introduce machine learning algorithms to monitor transaction behavior in real time, and immediately generate warning information if behavior that meets the trigger conditions is found. Once the system detects an event that meets the trigger conditions (such as abnormal transactions), the corresponding business process is immediately executed. For example, further investigation and analysis of accounts A and B is initiated. After triggering, the workflow is automatically generated. Request the risk control team to review the suspected account and provide detailed traceability information and identification results. Run the intelligent business analysis module to evaluate the transaction pattern between accounts and generate a behavior pattern analysis report. Automatically send alerts to relevant departments (such as customer service, risk control team) to notify transaction information with potential risks. According to the results of the intelligent analysis, if the analysis shows high risk, manual intervention may be required (such as freezing accounts, suspending transactions, etc.). If the risk is controllable, continue to monitor subsequent transactions. Update the processing results of this event and feedback them to the system, optimize the intelligent analysis algorithm and trigger mechanism, so that it can more accurately identify risks in the future.

[0097] For example, suppose that in a financial platform, the system detects that account A transfers money to multiple accounts in a short period of time, and the transfer amount is large. At this time, the following specific business process is triggered: The threshold trigger condition set by the system shows that the transaction behavior of account A is abnormal. The risk control system automatically triggers the investigation process and requests an audit of account A's transactions. Perform intelligent analysis to provide a comprehensive report on the transaction pattern, transaction history and identity authentication between account A and its target account. Send an alert to the risk control team, requesting further review of account A and proposing corresponding countermeasures. Through the above steps, it can be clearly seen how to trigger the target data business for the object to be identified and complete the intelligent business analysis based on the target identity information and target traceability information. Such a systematic process not only improves the efficiency of risk identification, but also enhances the ability of enterprises to respond to potential fraud or abnormal behavior. Ensuring the security of financial transactions and user funds is an important guarantee for the successful operation of enterprises.

[0098] Further optionally, after 104, intelligent business analysis optimization can be effectively improved through visualization tools and enhanced learning feedback. In the financial risk management system, enterprises can use data visualization tools such as Tableau, Power BI or QlikView to integrate them with existing data warehouses and data analysis modules to ensure smooth data automation processing. This integration enables the system to directly generate visual charts and provide intuitive data insights. For example, design a user-friendly dashboard that contains a variety of key charts, such as risk trend charts, transaction flow charts, and anomaly detection charts. These charts not only focus on user experience in design, but also ensure that data can be updated in real time, so that decision makers can instantly view the latest analysis results. Through such a visual dashboard, users can monitor business conditions and effectively view daily transaction volumes, the number of abnormal transactions, and detailed information involving high-risk accounts. This information helps the risk control team to quickly identify potential risks and take corresponding actions immediately when necessary, such as further investigating suspicious transactions or restricting the transaction permissions of high-risk accounts. At the same time, in order to improve the performance of financial fraud detection models, enterprises can also design enhanced learning feedback mechanisms. By regularly analyzing historical fraud detection data and related processing results, enterprises can evaluate the advantages and disadvantages of their models in handling similar cases. This process can be achieved by establishing a feedback loop system, that is, inputting the collected new data and known fraud case information into the model so that it can continuously learn and adapt to new fraud patterns. In terms of model optimization, companies can automatically adjust the parameters of the model based on feedback results. This dynamic adjustment can significantly improve the accuracy of future detection. In addition, as fraud methods continue to change, companies need to continuously upgrade and improve detection algorithms based on new features identified. This flexibility enables the model to effectively respond to new fraud behaviors.

[0099] In summary, by combining visualization tools with enhanced learning feedback mechanisms, enterprises can not only improve the performance and accuracy of intelligent business analysis, but also achieve more efficient risk management and decision support in practical applications. This comprehensive approach enables financial enterprises to have stronger adaptability and competitiveness in the rapidly changing market environment.

[0100] Further optionally, taking the example of a medical institution, the security and compliance of protecting sensitive data is critical, especially when using a data warehouse to store and process patient information. To ensure the security of sensitive medical records, medical institutions can implement a series of data encryption and security audit measures.

[0101] First, in the implementation of data encryption, medical institutions choose to use standard encryption technology, such as AES-256. This algorithm encrypts sensitive medical records stored in the data warehouse, ensuring that even if unauthorized visitors obtain the data, they cannot interpret the information. In addition, during the data transmission process, the use of TLS (Transport Layer Security) technology can ensure the confidentiality and integrity of data when it is transmitted from one system to another. This practice is particularly important because medical institutions often need to share patient information, and TLS can prevent data leakage and tampering, so that data security during transmission can be guaranteed.

[0102] At the same time, the implementation of a comprehensive security audit mechanism is also an indispensable part. In the data warehouse, enable detailed audit logging functions to record all access and operation behaviors involving sensitive data. For example, the system will automatically record when, who and how a specific patient's medical record was accessed. This not only provides a basis for tracking potential illegal access, but also provides the necessary evidence for compliance audits. On this basis, medical institutions are also required to conduct regular security audits, regularly check whether data access permissions comply with established security policies, and evaluate whether operations comply with relevant laws and regulations (such as HIPAA). Through audits, medical institutions can promptly discover and deal with potential security issues to ensure that sensitive data is not abused or leaked.

[0103] Implementing a comprehensive security audit mechanism is a key step in protecting sensitive data in healthcare organizations. Here are the specific implementation steps to ensure that sensitive medical records in the data warehouse can be effectively monitored and managed. First, healthcare organizations need to enable detailed audit logging in the data warehouse system. This function should be able to capture and record all access and operation behaviors involving sensitive data.

[0104] Selecting an audit log system: Healthcare organizations should select an appropriate audit log system and ensure that it can be seamlessly integrated with the existing data warehouse.

[0105] Define the level of logging: Set specific logging standards, such as the time of access, the identity of the visitor (such as user ID), the specific type of operation (such as read, modify, or delete), and the result of the operation (success or failure).

[0106] Configure log retention policy: define the retention period, storage location, and protection measures for audit logs to ensure the security of the log data itself.

[0107] To ensure that only authorized personnel can access sensitive data, medical institutions need to establish a strict data access control mechanism. Role-based access control (RBAC) sets access rights according to the user's role and the responsibilities associated with it. Different roles should have different access levels to sensitive data. Establish a permission approval process, and all new data access requests must be reviewed and approved by relevant supervisors. Set up a regular permission review mechanism to ensure that the access rights of existing users still meet business needs and revoke permissions that are no longer needed in a timely manner. For example, medical institutions should develop regular security audit plans to verify whether data access rights and operations comply with established security policies. You can also determine the frequency of audits, such as monthly, quarterly, or annually, depending on the size of the organization and the complexity of data processing. Clarify the content of the audit, including access records, permission settings, data processing processes, etc., to ensure that all sensitive data operations are covered. And push it to the compliance team or an independent third-party audit agency to perform the audit, evaluate and check the audit logs, and analyze whether there are any violations of security policies. After the audit is completed, medical institutions need to conduct risk assessments and deal with potential problems. Conduct a detailed analysis of the audit results and generate an audit report to point out the security risks and potential risks found. Develop rectification plans for the problems found in the audit, such as strengthening training, revising security policies, or optimizing access control policies. Set up a tracking mechanism to ensure that the rectification measures are fully implemented and check their effectiveness again in the next audit. Finally, medical institutions should establish a continuous improvement mechanism to ensure that security audit work is always efficient and agile. Regularly evaluate and update the audit process to adapt to changes in laws, regulations and industry best practices. Regularly conduct security awareness training for employees to help them understand the importance of data security and the need to comply with relevant policies. Pay close attention to technological developments and upgrade data warehouses and audit systems when necessary to improve data protection capabilities. Through the above steps, medical institutions can not only implement effective security audit mechanisms, but also ensure the protection and compliance of sensitive data and improve the overall level of data security management.

[0108] Such comprehensive measures enable medical institutions to effectively protect sensitive data, reduce the risk of data leakage, and enhance patient trust, ensuring that they follow strict security and compliance standards while providing high-quality medical services. By implementing encryption and security audits in data warehouses, medical institutions not only protect their most important asset - sensitive patient data, but also enhance their own security management capabilities.

[0109] In some implementations, the intelligent business analysis and optimization device based on the multidimensional data warehouse can be deployed in a separate manner, referring to Figure 2 The intelligent business analysis optimization method based on multidimensional data warehouse provided in the embodiment of the present application can be based on Figure 2The illustrated embodiment shows an intelligent business analysis and optimization system based on a multidimensional data warehouse. The intelligent business analysis and optimization system may include a server 01 and a terminal device 02 .

[0110] The server 01 may be loaded with an intelligent business analysis and optimization device, in which an intelligent business analysis and optimization program may be deployed. The terminal device 02 may be an intelligent business analysis and optimization device, in which an intelligent business analysis and optimization model may be deployed, such as a dynamic data analysis model and a dynamic tracing model trained based on a machine learning method. The terminal device 02 may forward the received object to be processed to the server 01. The server 01 may obtain the data attribute information of the object to be identified; the object to be identified is data information from at least one data source in a multidimensional data source; the data attribute information at least includes: data content, data type, data source identifier, data source type, and data collection method of the object to be identified; the data attribute information is input into the dynamic data analysis model, and the data identity of the object to be identified is identified to obtain the target identity information of the object to be identified; based on the target identity information, the dynamic tracing model is used to trace the identity of the object to be identified to obtain the target tracing information of the object to be identified; based on the target identity information and the target tracing information, the target data service for the object to be identified is triggered, and then the trigger instruction or execution information is sent to the terminal device 02 to complete the intelligent business analysis of the object to be processed.

[0111] It should be noted that the computing device involved in the embodiments of the present application may be a server and / or a terminal device.

[0112] The server involved in the embodiments of the present application can be an independent physical server, a server cluster or a distributed system composed of multiple physical servers, or a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, CDN, as well as big data and artificial intelligence platforms.

[0113] In the embodiment of the present application, the implementation of dynamic data analysis and identity traceability processes in a multidimensional data warehouse can bring many beneficial effects. First, by comprehensively integrating the data attribute information of the object to be identified and inputting the dynamic data analysis model, the rich information in the multidimensional data source can be used to significantly improve the accuracy of identity recognition. The dynamic analysis model adapts to the ever-changing data characteristics and reduces the probability of misidentification and missed identification. Secondly, the use of a dynamic traceability model can effectively track the source and changes of data, clarify the reliability and change history of data, thereby improving data traceability capabilities and providing support for data supervision and compliance review. In addition, combining target identity information with target traceability information can promote intelligent business analysis, provide a basis for decision support strategies, such as personalized recommendations and risk assessment, and improve business operation efficiency. Dynamic traceability management can also enhance data security and transparency, ensure that the source of data is clear and traceable, enhance users' trust in data management, and promptly discover potential security risks and respond quickly. By integrating multidimensional data sources, the correlation and potential value between data can be mined to help companies conduct market analysis and consumer behavior analysis, so as to better meet user needs and optimize products and services. In addition, the dynamic data analysis and traceability model enables enterprises to adjust analysis and decision-making strategies in real time according to market changes, quickly respond to changes and seize market opportunities, and enhance competitive advantages. Finally, through automated data identification and traceability, human intervention is reduced, data processing efficiency is improved, and collaboration between departments is promoted. In summary, the dynamic data analysis optimization method based on the multidimensional data warehouse not only improves the data identification and traceability capabilities, but also provides strong support for intelligent business analysis, and is committed to achieving data-driven decision-making and promoting sustainable development and growth of the business.

[0114] In another embodiment of the present application, an intelligent business analysis and optimization device based on a multidimensional data warehouse is also provided. Figure 3 As shown, the device includes the following units:

[0115] The acquisition unit is configured to acquire data attribute information of the object to be identified for the object to be processed loaded in the data warehouse; the object to be identified is data information from at least one data source in the multidimensional data source; the data attribute information at least includes: data content, data type, data source identifier, data source type, and data collection method of the object to be identified;

[0116] an identification unit configured to input the data attribute information into a dynamic data analysis model, perform data identity identification on the object to be identified, and obtain target identity information of the object to be identified;

[0117] A tracing unit is configured to perform identity tracing on the object to be identified based on the target identity information by using a dynamic tracing model to obtain target tracing information of the object to be identified;

[0118] The trigger unit is configured to trigger the target data service for the object to be identified based on the target identity information and the target traceability information, so as to complete the intelligent service analysis of the object to be processed.

[0119] Further optionally, the dynamic data analysis model comprises at least: a first feature extraction layer, an analysis layer, a prediction layer, and an optimization layer;

[0120] The access permissions include at least storage permissions and equipment permission permissions.

[0121] The identification unit inputs the data attribute information into the dynamic data analysis model, performs data identity identification on the object to be identified, and obtains the target identity information of the object to be identified, and is specifically configured to: extract the data individualized attribute features corresponding to the object to be identified from the data attribute information through the first feature extraction layer;

[0122] By means of the analysis layer, key feature analysis is performed on the individualized attribute features of the data to obtain key attribute features corresponding to the individualized attribute features of the data;

[0123] Through the prediction layer, identity feature prediction is performed according to the key attribute features to obtain a probability distribution of candidate identities of the object to be identified;

[0124] Through the optimization layer, non-maximum suppression processing is adopted to select the target identity information with the highest identity probability value.

[0125] Further optionally, the probability distribution of the candidate identities of the object to be identified is expressed as follows:

[0126]

[0127] Among them, P(I j |x) represents the probability of the candidate identity of the jth object to be identified under the key attribute feature x. j , w k represents the first weight item corresponding to the key attribute feature x, b k represents the first bias item corresponding to the key attribute feature x, w j represents the second weight item corresponding to the key attribute feature x, b j Represents the second bias term corresponding to the key attribute feature x.

[0128] Further optionally, after the identification unit inputs the data attribute information into the dynamic data analysis model, performs data identity identification on the object to be identified, and obtains the target identity information of the object to be identified, it is further configured to:

[0129] Verify whether the target identity information is consistent with the reference identity information indicated by the preset verification information contained in the data attribute information; the preset verification information is at least one of the data type, data source identifier, data source type, and data collection method; the reference identity information is an identity type pre-set in a preset identity verification library and bound to the preset verification information;

[0130] If the target identity information is consistent with the reference identity information, it is determined that the target identity information is verified;

[0131] If the target identity information is inconsistent with the reference identity information, the step of inputting the data attribute information into the dynamic data analysis model, performing data identity recognition on the object to be identified, and obtaining the target identity information of the object to be identified is triggered and executed again.

[0132] Further optionally, after the identification unit inputs the data attribute information into the dynamic data analysis model, performs data identity identification on the object to be identified, and obtains the target identity information of the object to be identified, it is further specifically configured to:

[0133] Verify whether the object to be identified has storage permission for the target storage area; the verification of the storage permission includes at least data attributes, access level, time range, and source category; the source category includes at least sensitive data and non-sensitive data; the time range includes at least: data valid range, data invalid range, and data custom effective range;

[0134] If the object to be identified does not have the storage permission, then the process stops based on the target identity information, and the dynamic tracing model is used to perform identity tracing on the object to be identified to obtain the target tracing information of the object to be identified.

[0135] Further optionally, the dynamic tracing model at least includes: a second feature extraction layer, a path construction layer, a mapping layer, an adjustment layer, and a tracing layer;

[0136] The tracing unit, based on the target identity information, uses a dynamic tracing model to trace the identity of the object to be identified, and obtains the target tracing information of the object to be identified, and is configured as follows:

[0137] Through the feature extraction layer, according to the data flow mode corresponding to the target identity information, the feature extraction of the object to be identified is performed to obtain the data flow feature of the object to be identified; the data flow feature at least includes: data path feature, data conversion feature, and data operation behavior feature;

[0138] By means of the path construction layer, a path is set based on the data flow characteristics of the object to be identified, so as to obtain a first data flow path of the object to be identified;

[0139] Through the mapping layer, the data flow features of the object to be identified are split into flow path feature points, and the flow path feature points are mapped to the spherical feature space to obtain the flow high-dimensional features corresponding to the flow path feature points;

[0140] Obtaining the cosine distance measurement similarity of the flow high-dimensional features through the adjustment layer, and optimizing and correcting the flow high-dimensional features by using the cosine distance measurement similarity to obtain optimized flow high-dimensional features;

[0141] Through the traceability layer, the optimized high-dimensional flow features of the flow path feature points are input into the relationship classifier to obtain the relationship classification prediction distribution value between the flow path feature points, and the target traceability information of the object to be identified is generated based on the relationship classification prediction distribution value.

[0142] Further optionally, the cosine distance metric of the flow high-dimensional feature is expressed as the following expression:

[0143] T i =S(i)(c·||ω i ||-α)cos(m·β i +d)

[0144] Among them, T i represents the cosine distance measurement similarity of the i-th flow high-dimensional feature vector, S(i) represents the i-th flow high-dimensional feature vector, α represents the cosine value amplification factor, c is the cosine value reduction factor, d is the cosine value balance adjustment factor, m is an adjustable hyperparameter, ω i represents the weight vector of the high-dimensional feature vector of the i-th flow, β i Represents the angle between the i-th flow high-dimensional feature vector and the weight vector.

[0145] In the embodiments of the present application, the data identification and traceability capabilities can be improved, and strong support can be provided for intelligent business analysis, which is committed to achieving data-driven decision-making.

[0146] In another embodiment of the present application, an electronic device is provided, comprising: a processor, a communication interface, a memory and a communication bus, wherein the processor, the communication interface and the memory communicate with each other via the communication bus;

[0147] Memory, used to store computer programs;

[0148] The processor is used to implement the intelligent business analysis optimization method based on a multidimensional data warehouse described in the method embodiment when executing the program stored in the memory.

[0149] The communication bus 1140 mentioned in the above electronic device may be a Peripheral Component Interconnect (PCI) bus or an Extended Industry Standard Architecture (EISA) bus, etc. The communication bus 1140 may be divided into an address bus, a data bus, a control bus, etc.

[0150] For ease of representation, Figure 4 Only one thick line is used in the diagram, but this does not mean that there is only one bus or only one type of bus.

[0151] The communication interface 1120 is used for communication between the above electronic device and other devices.

[0152] The memory 1130 may include a random access memory (RAM) or a non-volatile memory (non-volatile memory), such as at least one disk memory. Optionally, the memory may also be at least one storage device located away from the aforementioned processor.

[0153] The above-mentioned processor 1110 can be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), etc.; it can also be a digital signal processor (DSP), an application specific integrated circuit (ASIC), a field programmable gate array (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components.

[0154] Accordingly, an embodiment of the present application further provides a computer-readable storage medium storing a computer program, which, when executed, can implement each step that can be executed by an electronic device in the above method embodiment.

Claims

1. An intelligent business analysis and optimization method based on a multidimensional data warehouse, characterized in that: include: For the object to be processed loaded in the data warehouse, obtaining data attribute information of the object to be identified; The object to be identified is data information from at least one data source in a multidimensional data source; The data attribute information at least includes: data content, data type, data source identifier, data source type, and data collection method of the object to be identified; Inputting the data attribute information into a dynamic data analysis model, performing data identity recognition on the object to be identified, and obtaining target identity information of the object to be identified; Based on the target identity information, a dynamic tracing model is used to trace the identity of the object to be identified, so as to obtain the target tracing information of the object to be identified; Based on the target identity information and the target traceability information, a target data service for the object to be identified is triggered to complete intelligent service analysis of the object to be processed.

2. The intelligent business analysis and optimization method based on multidimensional data warehouse according to claim 1 is characterized in that: The dynamic data analysis model at least includes: a first feature extraction layer, an analysis layer, a prediction layer, and an optimization layer; The step of inputting the data attribute information into a dynamic data analysis model, performing data identity recognition on the object to be recognized, and obtaining target identity information of the object to be recognized includes: Extracting, through the first feature extraction layer, data individualized attribute features corresponding to the object to be identified from the data attribute information; By means of the analysis layer, key feature analysis is performed on the individualized attribute features of the data to obtain key attribute features corresponding to the individualized attribute features of the data; Through the prediction layer, identity feature prediction is performed according to the key attribute features to obtain a probability distribution of candidate identities of the object to be identified; Through the optimization layer, non-maximum suppression processing is adopted to select the target identity information with the highest identity probability value.

3. The intelligent business analysis and optimization method based on multidimensional data warehouse according to claim 2 is characterized in that: The probability distribution of the candidate identities of the object to be identified is expressed as follows: Among them, P(I j |x) represents the probability of the candidate identity of the jth object to be identified under the key attribute feature x. j , w k represents the first weight item corresponding to the key attribute feature x, b k represents the first bias item corresponding to the key attribute feature x, w j represents the second weight item corresponding to the key attribute feature x, b j Represents the second bias term corresponding to the key attribute feature x.

4. The intelligent business analysis and optimization method based on multidimensional data warehouse according to claim 2 is characterized in that: After inputting the data attribute information into the dynamic data analysis model, performing data identity recognition on the object to be recognized, and obtaining the target identity information of the object to be recognized, the method further includes: Verify whether the target identity information is consistent with the reference identity information indicated by the preset verification information contained in the data attribute information; the preset verification information is at least one of the data type, data source identifier, data source type, and data collection method; the reference identity information is an identity type pre-set in a preset identity verification library and bound to the preset verification information; If the target identity information is consistent with the reference identity information, it is determined that the target identity information is verified; If the target identity information is inconsistent with the reference identity information, the step of inputting the data attribute information into the dynamic data analysis model, performing data identity recognition on the object to be identified, and obtaining the target identity information of the object to be identified is triggered and executed again.

5. The intelligent business analysis and optimization method based on multidimensional data warehouse according to claim 2 is characterized in that: After inputting the data attribute information into the dynamic data analysis model, performing data identity recognition on the object to be recognized, and obtaining the target identity information of the object to be recognized, the method further includes: Verify whether the object to be identified has storage permission for the target storage area; the verification of the storage permission includes at least data attributes, access level, time range, and source category; the source category includes at least sensitive data and non-sensitive data; the time range includes at least: data valid range, data invalid range, and data custom effective range; If the object to be identified does not have the storage permission, then the process stops based on the target identity information, and the dynamic tracing model is used to perform identity tracing on the object to be identified to obtain the target tracing information of the object to be identified.

6. The intelligent business analysis and optimization method based on multidimensional data warehouse according to claim 1 is characterized in that: The dynamic tracing model at least includes: a second feature extraction layer, a path construction layer, a mapping layer, an adjustment layer, and a tracing layer; The step of tracing the identity of the object to be identified by using a dynamic tracing model based on the target identity information to obtain the target tracing information of the object to be identified includes: Through the feature extraction layer, according to the data flow mode corresponding to the target identity information, the feature extraction of the object to be identified is performed to obtain the data flow feature of the object to be identified; the data flow feature at least includes: data path feature, data conversion feature, and data operation behavior feature; By means of the path construction layer, a path is set based on the data flow characteristics of the object to be identified, so as to obtain a first data flow path of the object to be identified; Through the mapping layer, the data flow features of the object to be identified are split into flow path feature points, and the flow path feature points are mapped to the spherical feature space to obtain the flow high-dimensional features corresponding to the flow path feature points; Obtaining the cosine distance measurement similarity of the flow high-dimensional features through the adjustment layer, and optimizing and correcting the flow high-dimensional features by using the cosine distance measurement similarity to obtain optimized flow high-dimensional features; Through the traceability layer, the optimized high-dimensional flow features of the flow path feature points are input into the relationship classifier to obtain the relationship classification prediction distribution value between the flow path feature points, and the target traceability information of the object to be identified is generated based on the relationship classification prediction distribution value.

7. The intelligent business analysis and optimization method based on multidimensional data warehouse according to claim 6 is characterized in that: The cosine distance metric of the flow high-dimensional feature is expressed as follows: T i =S(i)(c·||ω i ||-α)cos(m·β i +d) Among them, T i represents the cosine distance measurement similarity of the i-th flow high-dimensional feature vector, S(i) represents the i-th flow high-dimensional feature vector, α represents the cosine value amplification factor, c is the cosine value reduction factor, d is the cosine value balance adjustment factor, m is an adjustable hyperparameter, ω i represents the weight vector of the high-dimensional feature vector of the i-th flow, β i Represents the angle between the i-th flow high-dimensional feature vector and the weight vector.

8. An intelligent business analysis and optimization device based on a multidimensional data warehouse, characterized in that: The device comprises: The acquisition unit is configured to acquire data attribute information of the object to be identified for the object to be processed loaded in the data warehouse; the object to be identified is data information from at least one data source in the multidimensional data source; the data attribute information at least includes: data content, data type, data source identifier, data source type, and data collection method of the object to be identified; an identification unit configured to input the data attribute information into a dynamic data analysis model, perform data identity identification on the object to be identified, and obtain target identity information of the object to be identified; A tracing unit is configured to perform identity tracing on the object to be identified based on the target identity information by using a dynamic tracing model to obtain target tracing information of the object to be identified; The trigger unit is configured to trigger the target data service for the object to be identified based on the target identity information and the target traceability information, so as to complete the intelligent service analysis of the object to be processed.

9. An electronic device, characterized in that: include: Memory for storing computer software programs; A processor is used to read and execute the computer software program, thereby implementing the intelligent business analysis and optimization method based on a multidimensional data warehouse as described in any one of claims 1 to 7.

10. A non-transitory computer-readable storage medium, characterized in that: The storage medium stores a computer software program, and when the computer software program is executed by the processor, the intelligent business analysis and optimization method based on a multidimensional data warehouse as described in any one of claims 1 to 7 is implemented.

Images