Search

CN-121996172-A - Management system for business data

CN121996172ACN 121996172 ACN121996172 ACN 121996172ACN-121996172-A

Abstract

The invention discloses a management system for business data, and relates to the technical field of data management. The management system for the service data comprises a data acquisition analysis module, an intelligent quality perception classification module, a storage drive evaluation module and a strategy fusion reasoning module, wherein the data acquisition analysis module is used for acquiring multi-source heterogeneous service data and extracting a service basic feature vector set, the intelligent quality perception classification module is used for inputting the service basic feature vector set into a pre-trained quality perception classification network to generate a quality confidence evaluation value, the storage drive evaluation module is used for extracting a storage drive evaluation value, and the strategy fusion reasoning module is used for carrying out fusion reasoning processing on the evaluation value to obtain an adaptive storage strategy matrix.

Inventors

  • Yin Shilian

Assignees

  • 北京天拓力行科技有限公司

Dates

Publication Date
20260508
Application Date
20260224

Claims (10)

  1. 1. A management system for business data, comprising: The data acquisition and analysis module is used for acquiring multi-source heterogeneous service data and extracting a service basic feature vector set of the multi-source heterogeneous service data; the intelligent quality perception classification module is used for inputting the service basic feature vector set into a pre-trained quality perception classification network, wherein the quality perception classification network comprises a feature extraction sub-network, a classification sub-network and a quality detection sub-network; the feature extraction sub-network performs feature transformation and depth mining processing on the basic feature vector to generate a service depth coding feature vector set corresponding to multi-source heterogeneous service data; The classifying sub-network performs cooperative classification processing on the service depth coding feature vector set, extracts a two-dimensional classifying tag set of the multi-source heterogeneous service data, and classifies the multi-source heterogeneous service data into a plurality of types of service data clusters; the quality detection sub-network performs quality diagnosis processing on the service data clusters to obtain quality confidence evaluation values corresponding to each type of service data clusters; The storage drive evaluation module is used for extracting storage drive evaluation values corresponding to each type of service data cluster based on each type of service data cluster; The policy fusion reasoning module is used for carrying out fusion reasoning processing on the storage drive evaluation value and the quality confidence evaluation value to obtain an adaptive storage policy matrix corresponding to each type of service data cluster; And the self-adaptive storage executing module is used for taking corresponding storage management measures for each type of service data cluster based on the self-adaptive storage strategy matrix.
  2. 2. The management system for service data according to claim 1, wherein the specific step of extracting the service base feature vector is as follows: Preprocessing multi-source heterogeneous service data; Performing feature construction processing based on the preprocessed multi-source heterogeneous service data to generate an original feature set; And carrying out standardization processing on the original feature set, and splicing the original feature set into a service basic feature vector set according to a preset sequence.
  3. 3. The management system for service data according to claim 1, wherein the feature extraction sub-network comprises a feature splitting layer, an embedding layer, a feature fusion layer, and a depth coding layer, and the specific steps of generating the service depth coding feature vector set are as follows: in the feature splitting layer, the basic feature vector set is separated to generate a service discrete classification feature set and a service numerical value feature set corresponding to multi-source heterogeneous service data; In the embedding layer, carrying out dense coding processing on the service discrete classification feature vector set to generate a discrete feature embedding vector set corresponding to multi-source heterogeneous service data; In the feature fusion layer, the discrete feature embedded vector set and the business continuous numerical value feature vector set are spliced and transformed to generate a fusion enhancement feature vector set corresponding to multi-source heterogeneous business data; And in the depth coding layer, carrying out depth mining mapping processing on the fusion enhancement feature vector set, and outputting the service depth coding feature vector set.
  4. 4. A management system for traffic data according to claim 3, characterized in that the specific step of generating said set of fusion enhanced feature vectors is as follows: performing dimension splicing processing on the discrete feature embedded vector set and the service continuous numerical value feature vector set to generate a service fusion feature vector set corresponding to multi-source heterogeneous service data; And carrying out transformation enhancement processing on the service fusion characteristic vector set to generate the fusion enhancement characteristic vector set.
  5. 5. The management system for service data according to claim 1, wherein the classifying sub-network comprises a feature dimension reduction layer, a label distribution layer and a classifying output layer, and the specific steps of generating several types of service data clusters are as follows: in the feature dimension reduction layer, performing dimension compression processing on the service depth coding feature vector set to generate a two-dimensional feature vector set corresponding to multi-source heterogeneous service data; In the label distribution layer, performing discretization label distribution processing on the two-dimensional feature vector set to generate a two-dimensional classification label set corresponding to multi-source heterogeneous service data; And in the classifying and outputting layer, classifying and processing the multi-source heterogeneous service data based on the two-dimensional classifying tag set, and outputting a plurality of types of service data clusters.
  6. 6. The system for managing business data according to claim 5, wherein the specific steps of generating the two-dimensional classification tag set are as follows: normalizing the two-dimensional feature vector set; based on the normalized two-dimensional feature vector set, carrying out attribution judgment processing by combining with a preset service classification granularity, and generating a space attribution mapping set corresponding to multi-source heterogeneous service data; And carrying out label distribution processing on the two-dimensional feature vector set based on the space attribution mapping set to generate the two-dimensional classification label set.
  7. 7. The system for managing service data according to claim 1, wherein the quality detection sub-network comprises a type-aware diagnosis layer, a quality adaptation layer, and a confidence output layer, and the specific steps of obtaining quality confidence assessment values corresponding to each type of service data clusters are as follows: In the type-aware diagnosis layer, each type of service data cluster is subjected to matching processing, and a diagnosis configuration vector corresponding to the corresponding service data cluster is generated; in the quality adaptation layer, carrying out directional extraction processing on the diagnosis configuration vector to generate a quality adaptation set corresponding to a corresponding service data cluster; And in the confidence output layer, carrying out fusion processing on the quality adaptation set, and outputting the quality confidence evaluation value.
  8. 8. The management system for service data according to claim 1, wherein the specific step of extracting the storage drive evaluation value is as follows: extracting a storage demand evaluation set corresponding to each type of service data cluster based on each type of service data cluster; And carrying out comprehensive processing on the storage demand evaluation set to generate storage drive evaluation values corresponding to each type of service data cluster.
  9. 9. The management system for service data according to claim 1, wherein the specific steps of obtaining the adaptive storage policy matrix are as follows: constructing a policy library containing a plurality of candidate storage policies, and generating a corresponding policy feature vector for each candidate storage policy; Based on each type of service data cluster, extracting a classification label corresponding to each type of service data cluster; Constructing a storage strategy demand vector corresponding to each type of service data cluster based on the quality confidence evaluation value, the storage drive evaluation value and the classification label; Screening candidate storage strategies in the strategy library based on the classification labels and the quality confidence evaluation values to obtain candidate strategy subsets corresponding to each type of service data clusters; and carrying out policy optimization processing based on the storage policy demand vector and the candidate policy subset, and constructing the self-adaptive storage policy matrix.
  10. 10. The management system for service data according to claim 9, wherein the specific steps of the policy preference process are as follows: analyzing the matching distance between the storage strategy demand vector and the strategy feature vector of each candidate storage strategy in the candidate strategy subsets, and sorting the candidate storage strategy subsets based on the matching distance; screening a pareto optimal non-dominant strategy solution set from the candidate strategy subsets after sorting based on a multi-objective optimization method; And determining a first selected storage strategy and a second selected storage strategy corresponding to each type of business data cluster from the non-dominant strategy solution set based on a preset rule so as to generate the self-adaptive storage strategy matrix.

Description

Management system for business data Technical Field The invention relates to the technical field of data management, in particular to a management system for business data. Background The business data management is a technical system which is specially responsible for carrying out persistence preservation, organization, maintenance, life cycle management and access support on various business data generated in the operation process in an enterprise information technology architecture, and has the core tasks of realizing efficient, economic and reliable utilization of storage resources on the premise of meeting continuous and compliance requirements of the business, and relating to the operation cost, data asset value and response capability of the enterprise. With the deep digital transformation of enterprises, business data has the trend of explosive growth and high complexity, and the data generally has multi-source heterogeneous characteristics, wherein multi-source means that the data is generated and stored in a plurality of independent systems such as customer relationship management, enterprise resource planning, supply chain management, log platform and the like, and heterogeneous data is represented by structural diversity such as regular relational database tables, semi-structured documents, unstructured text logs and multimedia files, and differences in business attributes such as different value densities, access modes, update frequencies and compliance requirements. Based on the above scheme, the prior art has the defects that the prior art is difficult to accurately capture the business value density and structural characteristics of data, the data cluster division is serious, accurate basis is difficult to provide for differentiated storage strategies, the influence of data quality on storage strategy adaptation is easy to ignore, the diagnosis and quantitative evaluation of the quality state of the data clusters are lacking, the mismatch problems of insufficient high-quality data storage resources and excessive occupation of resources by low-quality data are easy to occur, the storage strategies are matched with unfused storage driving requirements and quality confidence evaluation values for collaborative reasoning, the mutual constraint targets of storage cost, access performance and data reliability are difficult to balance, the storage resource waste or business access performance bottleneck is easy to be caused, and therefore the storage requirements of multiple types of business data clusters are difficult to adapt, and the accuracy and the high efficiency of business data storage management are further reduced. Disclosure of Invention Aiming at the defects of the prior art, the invention provides a management system for business data, which solves the problem that the prior art is difficult to accurately match data characteristics and quality and the management is low-efficient due to mismatching of storage strategies. The management system for service data comprises a data acquisition and analysis module, a data acquisition and analysis module and a service basic feature vector set, wherein the data acquisition and analysis module is used for acquiring multi-source heterogeneous service data and extracting the service basic feature vector set of the multi-source heterogeneous service data; The intelligent quality perception classifying module is used for inputting the service basic feature vector set into a pre-trained quality perception classifying network, wherein the quality perception classifying network comprises a feature extraction sub-network, a classifying sub-network and a quality detection sub-network, wherein the feature extraction sub-network performs feature transformation and depth mining processing on the basic feature vector to generate a service depth coding feature vector set corresponding to multi-source heterogeneous service data; The system comprises a storage drive evaluation module, a strategy fusion reasoning module and an adaptive storage execution module, wherein the storage drive evaluation module is used for extracting storage drive evaluation values corresponding to each type of service data cluster based on each type of service data cluster, the strategy fusion reasoning module is used for carrying out fusion reasoning processing on the storage drive evaluation values and the quality confidence evaluation values to obtain an adaptive storage strategy matrix corresponding to each type of service data cluster, and the adaptive storage execution module is used for taking corresponding storage management measures for each type of service data cluster based on the adaptive storage strategy matrix. The method comprises the specific steps of extracting service basic feature vectors, namely preprocessing multi-source heterogeneous service data, carrying out feature construction processing on the basis of the preprocessed multi-source heterogeneous service data