Digital economic data intelligent management method and system

Abstract

The invention relates to the technical field of data management, and provides a digital economic data intelligent management method and system, and the method comprises the steps: carrying out the quality optimization processing of digital economic data, obtaining optimized data, and analyzing a data descriptor corresponding to data metadata; identifying a data identifier corresponding to the optimized data, analyzing a data index corresponding to the optimized data, analyzing a data coupling relationship among data in the optimized data, and performing clustering processing on the optimized data to obtain clustered data; detecting a data time sequence corresponding to the cluster data, constructing a visual chart of the cluster data, analyzing a data trend corresponding to the cluster data, extracting data characteristics of the cluster data, and monitoring a data change direction corresponding to the cluster data in real time; and constructing a management strategy corresponding to the cluster data, calculating the reliability of the management strategy, and taking the management strategy corresponding to the reliability as a final management strategy of the cluster data. The invention aims to improve the efficiency of intelligent management of digital economic data.

Description

Technical Field

[0001] The present invention relates to the field of data management technology, and in particular to a method and system for intelligent management of digital economic data. Background Art

[0002] Digital economic data refers to various data generated, collected, stored, processed and analyzed in digital economic activities, such as enterprise data. Data generated during enterprise operations, such as sales data, financial data, and production data. If the amount of data is too large, the processing speed of subsequent data will slow down, making data management difficult. Therefore, digital economic data needs to be managed.

[0003] The existing digital economy data management adopts the data warehouse method. The steps of this method are: clarify the goals and requirements of the data warehouse, design a suitable data model based on the goals and requirements, determine the source of the data, filter and process the data, load the filtered data into the data warehouse, select a suitable data storage method, provide data access and query interfaces, and regularly update the data to ensure data security and backup. Since the data is generated in real time, this method is difficult to adapt to the ever-changing data needs, which in turn leads to reduced management efficiency of digital economy data. Summary of the invention

[0004] The present invention provides a method and system for intelligent management of digital economic data, the main purpose of which is to improve the efficiency of intelligent management of digital economic data.

[0005] To achieve the above-mentioned purpose, the present invention provides a method for intelligent management of digital economic data, comprising:

[0006] Collecting digital economic data to be managed, performing quality optimization processing on the digital economic data to obtain optimized data, extracting data metadata corresponding to the optimized data, and analyzing data descriptors corresponding to the data metadata;

[0007] Identify a data identifier corresponding to the optimized data, analyze a data indicator corresponding to the optimized data according to the data identifier, analyze a data coupling relationship between data in the optimized data in combination with the data descriptor and the data indicator, and perform cluster processing on the optimized data according to the data coupling relationship to obtain cluster data, wherein the combining the data descriptor and the data indicator to analyze the data coupling relationship between data in the optimized data includes:

[0008] The indicator similarity between the data indicators is calculated by the following formula:

[0009]

[0010] Among them, Q represents the indicator similarity between data indicators, M i Indicates the indicator vector value corresponding to the i-th indicator in the data indicator, M i+1 Represents the indicator vector value corresponding to the i+1th indicator in the data indicator, and i and i+1 represent the sequence numbers corresponding to the data indicator;

[0011] Determining, according to the indicator similarity, related indicators among the data indicators;

[0012] Analyze the descriptor elements corresponding to the data descriptor, and construct an indicator element matrix corresponding to the associated indicator according to the descriptor elements;

[0013] Analyzing the indicator dependency relationship between the associated indicators according to the indicator element matrix;

[0014] Determining the data coupling relationship between the data in the optimized data according to the indicator dependency relationship;

[0015] Detecting the data time series corresponding to the cluster data, constructing a visualization chart of the cluster data based on the data time series, analyzing the data trend corresponding to the cluster data according to the visualization chart, and extracting the data features of the cluster data, and monitoring the data change direction corresponding to the cluster data in real time according to the data trend and the data features;

[0016] Based on the data change direction, the data demand corresponding to the cluster data is determined to construct a management strategy corresponding to the cluster data, and the reliability of the management strategy is calculated. When the reliability is greater than a preset threshold, the management strategy corresponding to the reliability is used as the final management strategy for the cluster data.

[0017] Optionally, the performing quality optimization processing on the digital economic data to obtain optimized data includes:

[0018] Performing data cleaning on the digital economic data to obtain cleaned data;

[0019] Performing noise reduction processing on the target data to obtain noise-reduced data;

[0020] Calculating the data variance value between the noise reduction data;

[0021] According to the data variance value, discrete data in the noise reduction data is eliminated to obtain optimized data.

[0022] Optionally, analyzing the data descriptor corresponding to the data metadata includes:

[0023] Mining the attribute information corresponding to the data metadata, and calculating the information entropy corresponding to the attribute information;

[0024] Normalizing the information entropy to obtain normalized information entropy;

[0025] Determining the information weight corresponding to the attribute information according to the normalized information entropy, and analyzing the information category corresponding to the attribute information;

[0026] In combination with the information category and the information weight, key attribute information is screened out from the attribute information;

[0027] A data descriptor corresponding to the data metadata is generated according to the key attribute information.

[0028] Optionally, analyzing a data indicator corresponding to the optimization data according to the data identifier includes:

[0029] Identify the identifier text in the data identifier, perform word segmentation processing on the identifier text, and obtain identifier segmentation;

[0030] Extracting the identifier core word from the identifier segmentation, performing semantic analysis on the identifier core word, and obtaining the identifier semantics;

[0031] According to the identifier semantics, the data indicator corresponding to the optimization data is determined from a preset indicator set table.

[0032] Optionally, the analyzing the descriptor element corresponding to the data descriptor includes:

[0033] Extracting constituent elements from the data descriptor, and normalizing the constituent elements to obtain normalized elements;

[0034] Calculating the element criticality corresponding to the normalized element, and setting the element sequence corresponding to the normalized element based on the element criticality;

[0035] Clustering the constituent elements to obtain cluster elements, and extracting core elements from the cluster elements based on the element sequence;

[0036] According to the core element, a descriptor element corresponding to the data descriptor is obtained.

[0037] Optionally, analyzing the indicator dependency relationship between the associated indicators according to the indicator element matrix includes:

[0038] Performing vectorization processing on each element in the indicator element matrix to obtain an element vector, and marking the element position corresponding to each element in the indicator element matrix;

[0039] According to the element vector and the element position, the element correlation coefficient between each element in the indicator element matrix is ​​calculated by the following formula:

[0040]

[0041] Among them, A represents the element correlation coefficient between each element in the indicator element matrix, B n Represents the element vector corresponding to the element at the nth element position in the Bth indicator element matrix, represents the matrix vector mean corresponding to the Bth indicator element matrix, B n+1 Represents the element vector corresponding to the element at the n+1th element position in the Bth indicator element matrix, where n and n+1 both represent element positions, and q represents the total number of element positions.

[0042] According to the element correlation coefficient, the indicator dependency relationship between the associated indicators is determined.

[0043] Optionally, extracting data features of the cluster data includes:

[0044] Analyzing the data format corresponding to the cluster data, and extracting initial data features corresponding to the cluster data based on the data format;

[0045] Calculate the characteristic Gini index corresponding to the initial data feature;

[0046] The initial data features are screened according to the characteristic Gini index to obtain the data features of the cluster data.

[0047] Optionally, calculating the characteristic Gini index corresponding to the initial data feature includes:

[0048] Performing standardization processing on the initial data features to obtain standard data features;

[0049] Quantifying the standard data features to obtain quantized data features;

[0050] Calculating a feature ratio corresponding to the quantitative data feature;

[0051] Combined with the feature ratio, the feature Gini coefficient corresponding to the standard data feature is calculated by the following formula:

[0052]

[0053] Among them, F represents the characteristic Gini coefficient corresponding to the standard data feature, j represents the sequence number corresponding to the standard data feature, r represents the total number of standard data features, and H j Indicates the feature ratio corresponding to the jth feature in the standard data features.

[0054] Optionally, calculating the reliability of the management strategy includes:

[0055] Analyze the policy indicators in the management policy, and schedule the historical indicator data corresponding to the policy indicators;

[0056] Calculate the indicator failure rate corresponding to the strategy indicator according to the indicator historical data;

[0057] Calculate the indicator priority corresponding to the strategy indicator;

[0058] Calculate the policy failure rate corresponding to the management policy by combining the indicator priority and the indicator failure rate;

[0059] The reliability of the management strategy is calculated according to the strategy failure rate.

[0060] A digital economic data intelligent management system, characterized in that the system comprises:

[0061] A descriptor analysis module, used to collect digital economic data to be managed, perform quality optimization processing on the digital economic data to obtain optimized data, extract data metadata corresponding to the optimized data, and analyze data descriptors corresponding to the data metadata;

[0062] A data clustering module is used to identify a data identifier corresponding to the optimized data, analyze a data indicator corresponding to the optimized data according to the data identifier, analyze a data coupling relationship between data in the optimized data in combination with the data descriptor and the data indicator, and perform clustering processing on the optimized data according to the data coupling relationship to obtain cluster data, wherein the combining the data descriptor and the data indicator to analyze the data coupling relationship between data in the optimized data includes:

[0063] The indicator similarity between the data indicators is calculated by the following formula:

[0064]

[0065] Among them, Q represents the indicator similarity between data indicators, M i Indicates the indicator vector value corresponding to the i-th indicator in the data indicator, M i+1 Represents the indicator vector value corresponding to the i+1th indicator in the data indicator, and i and i+1 represent the sequence numbers corresponding to the data indicator;

[0066] Determining, according to the indicator similarity, related indicators among the data indicators;

[0067] Analyze the descriptor elements corresponding to the data descriptor, and construct an indicator element matrix corresponding to the associated indicator according to the descriptor elements;

[0068] Analyzing the indicator dependency relationship between the associated indicators according to the indicator element matrix;

[0069] Determining the data coupling relationship between the data in the optimized data according to the indicator dependency relationship;

[0070] A feature extraction module is used to detect the data time series corresponding to the cluster data, construct a visualization chart of the cluster data based on the data time series, analyze the data trend corresponding to the cluster data according to the visualization chart, and extract the data features of the cluster data, and monitor the data change direction corresponding to the cluster data in real time according to the data trend and the data features;

[0071] A strategy construction module is used to determine the data requirements corresponding to the cluster data based on the direction of data change, to construct a management strategy corresponding to the cluster data, and to calculate the reliability of the management strategy. When the reliability is greater than a preset threshold, the management strategy corresponding to the reliability is used as the final management strategy for the cluster data.

[0072] The present invention can improve the quality and reliability of the digital economic data by performing quality optimization processing on the digital economic data, and reduce the complexity of subsequent data processing. The present invention analyzes the data indicators corresponding to the optimized data according to the data identifier, and then understands the data role corresponding to the optimized data, and improves the accuracy of the subsequent data coupling relationship analysis between the data in the optimized data. The present invention analyzes the data trend corresponding to the cluster data according to the visualization chart, and can understand the data trend corresponding to the cluster data, and extract the data characteristics of the cluster data, and can obtain the data representation corresponding to the cluster data. By combining the subsequent improvement of the monitoring accuracy of the subsequent data change direction, the present invention determines the data demand corresponding to the cluster data based on the data change direction to construct the management strategy corresponding to the cluster data, and can obtain the management method corresponding to the cluster data. By calculating the reliability of the management strategy, the reliability corresponding to the management strategy can be understood, and then it is convenient to determine the subsequent final management strategy to improve the management efficiency of the cluster data. Therefore, a digital economic data intelligent management method provided by an embodiment of the present invention can improve the efficiency of digital economic data intelligent management. BRIEF DESCRIPTION OF THE DRAWINGS

[0073] Figure 1 A flowchart of a method for intelligent management of digital economic data provided by one embodiment of the present invention;

[0074] Figure 2 A functional module diagram of a digital economic data intelligent management system provided for one embodiment of the present invention.

[0075] The realization of the purpose, functional features and advantages of the present invention will be further explained in conjunction with embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0076] It should be understood that the specific embodiments described herein are only used to explain the present invention, and are not used to limit the present invention.

[0077] The embodiment of the present application provides a method for intelligent management of digital economic data. In the embodiment of the present application, the execution subject of the method for intelligent management of digital economic data includes but is not limited to at least one of the electronic devices such as a server, a terminal, etc. that can be configured to execute the method provided by the embodiment of the present application. In other words, the method for intelligent management of digital economic data can be executed by software or hardware installed on a terminal device or a server device, and the software can be a blockchain platform. The server includes but is not limited to: a single server, a server cluster, a cloud server or a cloud server cluster, etc. The server can be an independent server, or it can be 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, content delivery networks (Content Delivery Network, CDN), and big data and artificial intelligence platforms.

[0078] Reference Figure 1 FIG. 1 is a flow chart of a method for intelligent management of digital economic data provided by an embodiment of the present invention. In this embodiment, the method for intelligent management of digital economic data includes steps S1 to S4.

[0079] S1. Collect the digital economic data to be managed, perform quality optimization processing on the digital economic data to obtain optimized data, extract data metadata corresponding to the optimized data, and analyze the data descriptors corresponding to the data metadata.

[0080] The present invention can improve the quality and reliability of the digital economic data and reduce the complexity of subsequent data processing by performing quality optimization processing on the digital economic data. The data economic data refers to the digital economy in the economic concept, which is the relevant data that humans use to guide and realize the rapid optimization allocation and regeneration of resources and achieve high-quality economic development through the identification, selection, filtering, storage and use of big data (digital knowledge and information). The optimized data is the data obtained after processing the invalid and abnormal data in the digital economic data.

[0081] As an embodiment of the present invention, the quality optimization processing of the digital economic data to obtain optimized data includes: performing data cleaning processing on the digital economic data to obtain cleaned data, performing noise reduction processing on the target data to obtain noise-reduced data, calculating the data variance value between the noise-reduced data, and according to the data variance value, eliminating discrete data in the noise-reduced data to obtain optimized data.

[0082] Among them, the cleaned data is the data obtained after removing missing data and duplicate data in the digital economic data, the denoised data is the data obtained after removing noise interference in the target data, and the data variance value indicates the degree of discreteness between the denoised data.

[0083] Furthermore, the data cleaning processing of the digital economic data can be achieved through the box plot method; the noise reduction processing of the target data can be achieved through a low-pass filter; the data variance value between the noise reduction data can be achieved through a variance calculator; the step of eliminating discrete data in the noise reduction data is: according to the data variance value, calculate the variance mean corresponding to the noise reduction data, compare the variance mean with the data variance value, if the data variance value is greater than the variance mean, the corresponding data is discrete data, and the discrete data is eliminated to obtain optimized data.

[0084] The present invention can obtain data description information corresponding to the optimized data by extracting data metadata corresponding to the optimized data, and can obtain features corresponding to the data metadata by analyzing data descriptors corresponding to the data metadata, thereby providing a basis for subsequent analysis of the data coupling relationship between data in the optimized data, wherein the data metadata is the description information corresponding to the optimized data, and the data descriptor is the specific feature or attribute corresponding to the data metadata. Furthermore, the extraction of the data metadata corresponding to the optimized data can be achieved through data governance tools, such as Collibra tools, which are tools focusing on data governance and metadata management.

[0085] As an embodiment of the present invention, the analysis of the data descriptor corresponding to the data metadata includes: mining the attribute information corresponding to the data metadata, calculating the information entropy corresponding to the attribute information, normalizing the information entropy to obtain normalized information entropy, determining the information weight corresponding to the attribute information based on the normalized information entropy, and analyzing the information category corresponding to the attribute information, combining the information category and the information weight, screening out key attribute information from the attribute information, and generating the data descriptor corresponding to the data metadata based on the key attribute information.

[0086] Among them, the information attribute is the data property corresponding to the data metadata, the information entropy represents the uncertainty corresponding to the attribute information, the larger the entropy value, the higher the information uncertainty, which means that the information is more important, the normalized information entropy is the result of the normalization processing of the information entropy, which facilitates the comparison between information entropies, and the information category represents the type of information corresponding to the attribute information.

[0087] Furthermore, the mining of attribute information corresponding to the data metadata can be achieved through a decision tree algorithm; the calculation of information entropy corresponding to the attribute information can be achieved through a Shannon entropy algorithm; the normalization processing of the information entropy can be achieved through a linear normalization method; the normalized information entropy can be directly used as the information weight corresponding to the attribute information, and the analysis of the information category corresponding to the attribute information can be achieved through natural language processing technology, such as grammatical analysis; key attribute information can be obtained by selecting the information with the largest information weight from the same information category; the generation step of the data descriptor corresponding to the data metadata is: analyzing the key attribute information to understand the meaning, relationship and characteristics of the information, and determining the variables that need to be included in the data descriptor based on the results of the analysis, such as the name, definition, data type, length, constraints, etc. of the data; based on the determined variables, constructing the structure of the data descriptor, usually a specific format or model can be used to represent the descriptor, and the key attribute information is filled into the descriptor structure to obtain the corresponding data descriptor.

[0088] S2. Identify the data identifier corresponding to the optimized data, analyze the data indicator corresponding to the optimized data according to the data identifier, analyze the data coupling relationship between the data in the optimized data in combination with the data descriptor and the data indicator, and cluster the optimized data according to the data coupling relationship to obtain cluster data.

[0089] The present invention analyzes the data indicators corresponding to the optimized data based on the data identifier, and then can understand the data functions corresponding to the optimized data, thereby improving the accuracy of subsequent data coupling relationship analysis between data in the optimized data, wherein the data identifier is a unique identifier corresponding to the optimized data, such as a description object of the data, and the data indicator is the data function corresponding to the optimized data. Furthermore, the identification of the data identifier corresponding to the optimized data can be achieved through a data analysis tool, such as an SPSS tool.

[0090] As an embodiment of the present invention, analyzing the data indicators corresponding to the optimized data based on the data identifier includes: identifying the identifier text in the data identifier, performing word segmentation processing on the identifier text to obtain identifier segmentation, extracting identifier core words in the identifier segmentation, performing semantic analysis on the identifier core words to obtain identifier semantics, and determining the data indicators corresponding to the optimized data from a preset indicator set table based on the identifier semantics.

[0091] Among them, the identifier text is the expression text corresponding to the data identifier, the identifier segmentation is the decomposition of the identifier text into single words or vocabulary items, the identifier core word is the key meaning and representative word in the identifier segmentation, the identifier semantics is the meaning and explanation corresponding to the identifier core word, and the preset indicator set table is a table composed of all indicators corresponding to the field of digital economic data.

[0092] Furthermore, the recognition of the identifier text in the data identifier can be achieved through OCR recognition technology; the word segmentation processing of the identifier text can be achieved through a word segmentation tool, and the word segmentation tool is compiled by a scripting language, such as JS scripting language; the identifier core words in the identifier word segmentation can be extracted by a machine learning algorithm, such as a hidden Markov model; the semantic analysis of the identifier core words can be obtained by semantic analysis; the identifier semantics can be compared with each indicator in a preset indicator set table, and if the same vocabulary appears, the indicator is used as the data indicator corresponding to the optimized data.

[0093] The present invention combines the data descriptor and the data indicator to analyze the data coupling relationship between the data in the optimized data, so as to understand the data correlation relationship between the optimized data, thereby providing a basis for subsequent clustering processing of the optimized data, wherein the data coupling relationship represents the data correlation relationship between the optimized data.

[0094] As an embodiment of the present invention, the combining the data descriptor and the data indicator to analyze the data coupling relationship between the data in the optimized data includes: calculating the indicator similarity between the data indicators by the following formula:

[0095]

[0096] Among them, Q represents the indicator similarity between data indicators, M i Indicates the indicator vector value corresponding to the i-th indicator in the data indicator, M i+1 Represents the indicator vector value corresponding to the i+1th indicator in the data indicator, i and i+1 represent the sequence number corresponding to the data indicator,

[0097] According to the indicator similarity, the associated indicators in the data indicators are determined, the descriptor elements corresponding to the data descriptors are analyzed, and according to the descriptor elements, an indicator element matrix corresponding to the associated indicators is constructed, and according to the indicator element matrix, the indicator dependency relationship between the associated indicators is analyzed, and according to the indicator dependency relationship, the data coupling relationship between the data in the optimized data is determined.

[0098] Among them, the indicator similarity represents the degree of similarity between the data indicators, the associated indicators are indicators with associated relationships among the data indicators, the descriptor elements are core components corresponding to the data descriptors, the indicator element matrix is ​​a square matrix composed of the descriptor elements corresponding to the associated indicators, and the indicator dependency is the corresponding relationship between the associated indicators.

[0099] Furthermore, the indicator similarity between the data indicators can be achieved through a cosine similarity algorithm; the associated indicators in the data indicators can be determined by comparing the indicator similarity with a preset threshold. If the indicator similarity is greater than the preset threshold, it means that the corresponding indicators are associated indicators; the indicator element matrix corresponding to the associated indicators can be achieved through a matrix function, such as a zero matrix function.

[0100] Optionally, as an optional embodiment of the present invention, the analysis of the descriptor elements corresponding to the data descriptor includes: extracting constituent elements in the data descriptor, normalizing the constituent elements to obtain normalized elements, calculating the element criticality corresponding to the normalized elements, setting an element sequence corresponding to the normalized elements based on the element criticality, and clustering the constituent elements to obtain cluster elements, extracting core elements from the cluster elements based on the element sequence, and obtaining the descriptor elements corresponding to the data descriptor according to the core elements.

[0101] Among them, the constituent elements are basic constituent elements in the data descriptor, the normalized elements are elements that are comparable after the constituent elements are scaled, the element criticality represents the criticality of the normalized elements, the element sequence is the serial number of the normalized elements, the clustering elements are the element sets obtained by clustering elements with the same attributes in the constituent elements, and the core elements are representative elements in the clustering elements. Furthermore, the extraction of constituent elements in the data descriptor can be achieved through grammatical analysis; the normalization of the constituent elements can be achieved through the minimum-maximum normalization method; the calculation of the element criticality corresponding to the normalized elements can be achieved through the principal component analysis method, and multiple normalized elements are converted into a set of new unrelated variables, called principal components, and the contribution rate of the principal components is calculated, and the contribution rate can be used as the element criticality; and the clustering of the constituent elements can be achieved through a clustering algorithm.

[0102] Optionally, as an optional embodiment of the present invention, analyzing the indicator dependency relationship between the associated indicators according to the indicator element matrix includes: vectorizing each element in the indicator element matrix to obtain an element vector, and marking the element position corresponding to each element in the indicator element matrix, and calculating the element correlation coefficient between each element in the indicator element matrix according to the element vector and the element position by the following formula:

[0103]

[0104] Among them, A represents the element correlation coefficient between each element in the indicator element matrix, B n Represents the element vector corresponding to the element at the nth element position in the Bth indicator element matrix, represents the matrix vector mean corresponding to the Bth indicator element matrix, B n+1 Represents the element vector corresponding to the element at the n+1th element position in the Bth indicator element matrix, where n and n+1 both represent element positions, and q represents the total number of element positions.

[0105] According to the element correlation coefficient, the indicator dependency relationship between the associated indicators is determined.

[0106] Among them, the element vector is the expression vector of each element in the indicator element matrix, the element position is the corresponding position of each element in the indicator element matrix in the matrix, and the element correlation coefficient represents the degree of association between each element in the indicator element matrix.

[0107] Furthermore, the vectorization processing of each element in the indicator element matrix can be implemented by the word2vec algorithm; the annotation of the element position corresponding to each element in the indicator element matrix can be implemented by an annotation tool, such as the VGGImageAnnotator tool; the indicator dependency relationship between the associated indicators can be determined according to the result of summing the element correlation coefficients. For example, if the result of summing the element correlation coefficients is a positive number, it indicates that the indicator dependency relationship between the associated indicators is a positive dependency relationship. If it is a negative number, it indicates that the indicator dependency relationship between the associated indicators is a negative dependency relationship.

[0108] The present invention performs cluster processing on the optimization data according to the data coupling relationship, and can group data with related relationships together to facilitate subsequent data trend analysis. Furthermore, the cluster processing of the optimization data can be achieved through a hierarchical clustering algorithm.

[0109] S3. Detect the data time series corresponding to the cluster data, construct a visualization chart of the cluster data based on the data time series, analyze the data trend corresponding to the cluster data according to the visualization chart, and extract data features of the cluster data; and monitor the data change direction corresponding to the cluster data in real time according to the data trend and the data features.

[0110] The present invention analyzes the data trend corresponding to the cluster data according to the visualization chart, can understand the data direction corresponding to the cluster data, and extracts the data features of the cluster data, can obtain the data representation corresponding to the cluster data, and improves the monitoring accuracy of the subsequent data change direction by combining with the subsequent, wherein the data time series is the time series corresponding to the cluster data, the visualization chart is the visualization table corresponding to the cluster data, the data trend is the description of the data direction corresponding to the cluster data, the data feature is the data representation corresponding to the cluster data, further, the data time series corresponding to the cluster data can be detected by the Kafka Streams tool; the construction of the visualization chart of the cluster data can be achieved by a drawing tool, such as the Visio drawing tool.

[0111] As an embodiment of the present invention, the extracting of data features of the cluster data includes: analyzing the data format corresponding to the cluster data, extracting initial data features corresponding to the cluster data in combination with the data format, calculating the characteristic Gini index corresponding to the initial data features, and screening the initial data features according to the characteristic Gini index to obtain the data features of the cluster data.

[0112] The data format is the data type corresponding to the cluster data, such as text or numerical data, the initial data features are all features in the cluster data, and the feature Gini index represents the feature importance corresponding to the initial data features.

[0113] Furthermore, the analysis of the data format corresponding to the cluster data can be achieved through a format analysis tool, such as a Python library; the feature extraction method corresponding to the cluster data is determined in combination with the data format, for example, for numerical features, statistical analysis methods (such as mean, standard deviation, etc.) can be used, and for categorical features, frequency analysis or pattern recognition methods can be used, and the initial data features corresponding to the cluster data are extracted according to the corresponding feature extraction method; the feature Gini indexes corresponding to features of the same type in the initial data features are compared, and the features with the smallest feature Gini index are screened out to obtain the data features of the cluster data.

[0114] Optionally, as an optional embodiment of the present invention, the calculating the characteristic Gini index corresponding to the initial data feature includes: standardizing the initial data feature to obtain a standard data feature, quantizing the standard data feature to obtain a quantized data feature, calculating a characteristic ratio corresponding to the quantized data feature, and combining the characteristic ratio to calculate the characteristic Gini coefficient corresponding to the standard data feature by the following formula:

[0115]

[0116] Among them, F represents the characteristic Gini coefficient corresponding to the standard data feature, j represents the sequence number corresponding to the standard data feature, r represents the total number of standard data features, and H j Indicates the feature ratio corresponding to the jth feature in the standard data features.

[0117] Among them, the standard data feature is the feature obtained after the differences between the initial data features are eliminated, the quantitative data feature is the expression value corresponding to the standard data feature, and the feature ratio is the ratio of each feature in the quantitative data feature.

[0118] Furthermore, the standardization processing of the initial data features can be achieved through the Z-Score standardization method; the quantization processing of the standard data features can be achieved through the one-hot encoding method; the feature ratio corresponding to the quantized data features can be obtained by calculating the number of times each feature appears in the quantized data features and the number corresponding to the total quantized data features.

[0119] The present invention monitors the data change direction corresponding to the cluster data in real time according to the data trend and the data feature, which is helpful to predict the future trend of the cluster data, and further can provide timely data support for decision-making. Furthermore, relevant features with the data change direction are screened out from the data features, and the feature effects corresponding to the relevant features are analyzed. The feature effects and the data trend are combined to monitor the data change direction corresponding to the cluster data. For example, if the feature effect is gain and the corresponding data trend is decreasing, it means that the corresponding cluster data changes in the direction of gain reduction. If the data trend is increasing, it means that the corresponding cluster data changes in the direction of gain increase.

[0120] S4. Based on the data change direction, determine the data demand corresponding to the cluster data to construct a management strategy corresponding to the cluster data, and calculate the reliability of the management strategy. When the reliability is greater than a preset threshold, use the management strategy corresponding to the reliability as the final management strategy for the cluster data.

[0121] The present invention determines the data demand corresponding to the cluster data based on the data change direction to construct a management strategy corresponding to the cluster data, and can obtain the management method corresponding to the cluster data. By calculating the reliability of the management strategy, the reliability corresponding to the management strategy can be understood, which is convenient for determining the subsequent final management strategy to improve the management efficiency of the cluster data, wherein the data demand is the demand for the cluster data to achieve the corresponding expectation or goal, such as the required associated data or the corresponding processing algorithm, the management strategy is the management method corresponding to the cluster data, and the preset threshold is the judgment standard of the reliability, which can be 0.8, and can also be set according to the actual application scenario. Optionally, the use scenario and related business objectives of the data are understood to determine the required data type and characteristics, and the specific data requirements are clarified according to the business objectives and the data change direction, such as specific indicators, dimensions, time periods, etc. based on the data change direction.

[0122] As an embodiment of the present invention, the calculation of the reliability of the management policy includes: analyzing the policy indicators in the management policy, scheduling the indicator historical data corresponding to the policy indicators, calculating the indicator failure rate corresponding to the policy indicators based on the indicator historical data, calculating the indicator priority corresponding to the policy indicators, calculating the policy failure rate corresponding to the management policy based on the indicator priority and the indicator failure rate, and calculating the reliability of the management policy based on the policy failure rate.

[0123] Among them, the strategic indicators are specific indicators used to measure the effectiveness and performance of management strategies, the indicator historical data are relevant data collected by the strategic indicators in the past period of time, which record the completion status of the indicator at different time points, the indicator failure rate represents the probability of failure corresponding to the strategic indicators, the indicator priority represents the importance of the strategic indicators in the management strategy, and the strategy failure rate represents the possibility of failure corresponding to the management strategy.

[0124] Furthermore, the analysis of the strategy indicators in the management strategy can be obtained by combining the process analysis of the management strategy, understanding the process involved in the management strategy, determining the key links and measurable indicators; the indicator historical data corresponding to the strategy indicators can be obtained by scheduling in the corresponding data warehouse; the calculation steps of the indicator failure rate corresponding to the strategy indicators are: filtering out the relevant execution data corresponding to all strategy indicators from the historical data, counting the number of failures of the strategy indicators, and counting the total number of executions of the strategy indicators in the historical data, dividing the number of failures by the total number of executions to obtain the indicator failure rate, the calculation of the indicator priority corresponding to the strategy indicator can be achieved through the hierarchical analysis method, comparing the indicators in the strategy indicators with each other, calculating the importance scores between each other, which can be obtained by calculating the degree of mutual influence, normalizing the importance scores to obtain the target score value, and using the target score value as the indicator priority, the strategy failure rate corresponding to the management strategy can be obtained by calculating the product of the indicator priority and the corresponding indicator failure rate, adding the products to obtain the strategy failure rate; the calculation of the reliability of the management strategy can be obtained by 1-the strategy failure rate.

[0125] The present invention can improve the quality and reliability of the digital economic data by performing quality optimization processing on the digital economic data, and reduce the complexity of subsequent data processing. The present invention analyzes the data indicators corresponding to the optimized data according to the data identifier, and then understands the data role corresponding to the optimized data, and improves the accuracy of the subsequent analysis of the data coupling relationship between the data in the optimized data. The present invention analyzes the data trend corresponding to the cluster data according to the visualization chart, and can understand the data trend corresponding to the cluster data, and extract the data characteristics of the cluster data, and can obtain the data representation corresponding to the cluster data. By combining the subsequent improvement of the monitoring accuracy of the subsequent data change direction, the present invention determines the data demand corresponding to the cluster data based on the data change direction to construct the management strategy corresponding to the cluster data, and can obtain the management method corresponding to the cluster data. By calculating the reliability of the management strategy, the reliability corresponding to the management strategy can be understood, and then it is convenient to determine the subsequent final management strategy to improve the management efficiency of the cluster data. Therefore, a digital economic data intelligent management method and system provided by an embodiment of the present invention can improve the efficiency of digital economic data intelligent management.

[0126] like Figure 2 The figure shows a functional module diagram of a digital economic data intelligent management system provided by one embodiment of the present invention.

[0127] The digital economic data intelligent management system 100 described in the present invention can be installed in an electronic device. According to the functions implemented, the digital economic data intelligent management system 100 can include a descriptor analysis module 101, a data clustering module 102, a feature extraction module 103 and a strategy construction module 104. The module described in the present invention can also be called a unit, which refers to a series of computer program segments that can be executed by an electronic device processor and can complete fixed functions, which are stored in the memory of the electronic device.

[0128] In this embodiment, the functions of each module / unit are as follows:

[0129] The descriptor analysis module 101 is used to collect the digital economic data to be managed, perform quality optimization processing on the digital economic data to obtain optimized data, extract data metadata corresponding to the optimized data, and analyze the data descriptors corresponding to the data metadata;

[0130] The data clustering module 102 is used to identify a data identifier corresponding to the optimized data, analyze the data index corresponding to the optimized data according to the data identifier, analyze the data coupling relationship between the data in the optimized data in combination with the data descriptor and the data index, and perform clustering processing on the optimized data according to the data coupling relationship to obtain cluster data, wherein the combining the data descriptor and the data index to analyze the data coupling relationship between the data in the optimized data includes:

[0131] The indicator similarity between the data indicators is calculated by the following formula:

[0132]

[0133] Among them, Q represents the indicator similarity between data indicators, M i Indicates the indicator vector value corresponding to the i-th indicator in the data indicator, M i+1 Represents the indicator vector value corresponding to the i+1th indicator in the data indicator, i and i+1 represent the sequence number corresponding to the data indicator,

[0134] Determining, according to the indicator similarity, related indicators among the data indicators;

[0135] Analyze the descriptor elements corresponding to the data descriptor, and construct an indicator element matrix corresponding to the associated indicator according to the descriptor elements;

[0136] Analyzing the indicator dependency relationship between the associated indicators according to the indicator element matrix;

[0137] Determining the data coupling relationship between the data in the optimized data according to the indicator dependency relationship;

[0138] The feature extraction module 103 is used to detect the data time series corresponding to the cluster data, construct a visualization chart of the cluster data based on the data time series, analyze the data trend corresponding to the cluster data according to the visualization chart, and extract the data features of the cluster data, and monitor the data change direction corresponding to the cluster data in real time according to the data trend and the data features;

[0139] The strategy construction module 104 is used to determine the data requirements corresponding to the cluster data based on the data change direction, so as to construct a management strategy corresponding to the cluster data and calculate the reliability of the management strategy. When the reliability is greater than a preset threshold, the management strategy corresponding to the reliability is used as the final management strategy for the cluster data.

[0140] In detail, each module described in the digital economic data intelligent management system 100 described in the embodiment of the present application adopts the same Figure 1 The technical means are the same as the digital economic data intelligent management method described in, and can produce the same technical effects, so I will not go into details here.

[0141] In the several embodiments provided by the present invention, it should be understood that the provided methods and systems can be implemented in other ways. For example, the method embodiments described above are only illustrative, for example, the division of the modules is only a logical function division, and there may be other division methods in actual implementation.

[0142] Finally, it should be noted that the above embodiments are only used to illustrate the technical solution of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solution of the present invention can be modified or replaced by equivalents without departing from the spirit and scope of the technical solution of the present invention.

Claims

1. A method for intelligent management of digital economic data, characterized in that: The method comprises: Collecting digital economic data to be managed, performing quality optimization processing on the digital economic data to obtain optimized data, extracting data metadata corresponding to the optimized data, and analyzing data descriptors corresponding to the data metadata; Identify a data identifier corresponding to the optimized data, analyze a data indicator corresponding to the optimized data according to the data identifier, analyze a data coupling relationship between data in the optimized data in combination with the data descriptor and the data indicator, and perform cluster processing on the optimized data according to the data coupling relationship to obtain cluster data, wherein the combining the data descriptor and the data indicator to analyze the data coupling relationship between data in the optimized data includes: The indicator similarity between the data indicators is calculated by the following formula: Among them, Q represents the indicator similarity between data indicators, M i Indicates the indicator vector value corresponding to the i-th indicator in the data indicator, M i+1 Represents the indicator vector value corresponding to the i+1th indicator in the data indicator, and i and i+1 represent the sequence numbers corresponding to the data indicator; Analyze the descriptor elements corresponding to the data descriptor, and construct an indicator element matrix corresponding to the associated indicator according to the descriptor elements; Analyzing the indicator dependency relationship between the associated indicators according to the indicator element matrix; Determining the data coupling relationship between the data in the optimized data according to the indicator dependency relationship; Detecting the data time series corresponding to the cluster data, constructing a visualization chart of the cluster data based on the data time series, analyzing the data trend corresponding to the cluster data according to the visualization chart, and extracting the data features of the cluster data, and monitoring the data change direction corresponding to the cluster data in real time according to the data trend and the data features; Based on the data change direction, the data demand corresponding to the cluster data is determined to construct a management strategy corresponding to the cluster data, and the reliability of the management strategy is calculated. When the reliability is greater than a preset threshold, the management strategy corresponding to the reliability is used as the final management strategy for the cluster data.

2. A digital economic data intelligent management method as claimed in claim 1, characterized in that: The performing quality optimization processing on the digital economic data to obtain optimized data includes: Performing data cleaning on the digital economic data to obtain cleaned data; Performing noise reduction processing on the target data to obtain noise-reduced data; Calculating the data variance value between the noise reduction data; According to the data variance value, discrete data in the noise reduction data is eliminated to obtain optimized data.

3. A digital economic data intelligent management method as claimed in claim 1, characterized in that: The analyzing the data descriptor corresponding to the data metadata includes: Mining the attribute information corresponding to the data metadata, and calculating the information entropy corresponding to the attribute information; Normalizing the information entropy to obtain normalized information entropy; Determining the information weight corresponding to the attribute information according to the normalized information entropy, and analyzing the information category corresponding to the attribute information; In combination with the information category and the information weight, key attribute information is screened out from the attribute information; A data descriptor corresponding to the data metadata is generated according to the key attribute information.

4. A digital economic data intelligent management method as claimed in claim 1, characterized in that: The step of analyzing the data indicator corresponding to the optimization data according to the data identifier includes: Identify the identifier text in the data identifier, perform word segmentation processing on the identifier text, and obtain identifier segmentation; Extracting the identifier core word from the identifier segmentation, performing semantic analysis on the identifier core word, and obtaining the identifier semantics; According to the identifier semantics, the data indicator corresponding to the optimization data is determined from a preset indicator set table.

5. A digital economic data intelligent management method as claimed in claim 1, characterized in that: The analyzing the descriptor element corresponding to the data descriptor includes: Extracting constituent elements from the data descriptor, and normalizing the constituent elements to obtain normalized elements; Calculating the element criticality corresponding to the normalized element, and setting the element sequence corresponding to the normalized element based on the element criticality; Clustering the constituent elements to obtain cluster elements, and extracting core elements from the cluster elements based on the element sequence; According to the core element, a descriptor element corresponding to the data descriptor is obtained.

6. A digital economic data intelligent management method as claimed in claim 1, characterized in that: Analyzing the indicator dependency relationship between the associated indicators according to the indicator element matrix includes: Performing vectorization processing on each element in the indicator element matrix to obtain an element vector, and marking the element position corresponding to each element in the indicator element matrix; According to the element vector and the element position, the element correlation coefficient between each element in the indicator element matrix is ​​calculated by the following formula: Among them, A represents the element correlation coefficient between each element in the indicator element matrix, B n Represents the element vector corresponding to the element at the nth element position in the Bth indicator element matrix, represents the matrix vector mean corresponding to the Bth indicator element matrix, B n+1 Represents the element vector corresponding to the element at the n+1th element position in the Bth indicator element matrix, where n and n+1 both represent element positions, and q represents the total number of element positions. According to the element correlation coefficient, the indicator dependency relationship between the associated indicators is determined.

7. A digital economic data intelligent management method as claimed in claim 1, characterized in that: The extracting the data features of the cluster data includes: Analyzing the data format corresponding to the cluster data, and extracting initial data features corresponding to the cluster data based on the data format; Calculate the characteristic Gini index corresponding to the initial data feature; The initial data features are screened according to the characteristic Gini index to obtain the data features of the cluster data.

8. A digital economic data intelligent management method as claimed in claim 7, characterized in that: The calculating the characteristic Gini index corresponding to the initial data feature includes: Performing standardization processing on the initial data features to obtain standard data features; Quantifying the standard data features to obtain quantized data features; Calculating a feature ratio corresponding to the quantitative data feature; Combined with the feature ratio, the feature Gini coefficient corresponding to the standard data feature is calculated by the following formula: Among them, F represents the characteristic Gini coefficient corresponding to the standard data feature, j represents the sequence number corresponding to the standard data feature, r represents the total number of standard data features, and H j Indicates the feature ratio corresponding to the jth feature in the standard data features.

9. A digital economic data intelligent management method as claimed in claim 1, characterized in that: The calculating the reliability of the management strategy includes: Analyze the policy indicators in the management policy, and schedule the historical indicator data corresponding to the policy indicators; Calculate the indicator failure rate corresponding to the strategy indicator according to the indicator historical data; Calculate the indicator priority corresponding to the strategy indicator; Calculate the policy failure rate corresponding to the management policy by combining the indicator priority and the indicator failure rate; The reliability of the management strategy is calculated according to the strategy failure rate.

10. A digital economic data intelligent management system, characterized in that: The system comprises: A descriptor analysis module, used to collect digital economic data to be managed, perform quality optimization processing on the digital economic data to obtain optimized data, extract data metadata corresponding to the optimized data, and analyze data descriptors corresponding to the data metadata; A data clustering module is used to identify a data identifier corresponding to the optimized data, analyze a data indicator corresponding to the optimized data according to the data identifier, analyze a data coupling relationship between data in the optimized data in combination with the data descriptor and the data indicator, and perform clustering processing on the optimized data according to the data coupling relationship to obtain cluster data, wherein the combining the data descriptor and the data indicator to analyze the data coupling relationship between data in the optimized data includes: The indicator similarity between the data indicators is calculated by the following formula: Among them, Q represents the indicator similarity between data indicators, M i Indicates the indicator vector value corresponding to the i-th indicator in the data indicator, M i+1 Represents the indicator vector value corresponding to the i+1th indicator in the data indicator, and i and i+1 represent the sequence numbers corresponding to the data indicator; Determining, according to the indicator similarity, related indicators among the data indicators; Analyze the descriptor elements corresponding to the data descriptor, and construct an indicator element matrix corresponding to the associated indicator according to the descriptor elements; Analyzing the indicator dependency relationship between the associated indicators according to the indicator element matrix; Determining the data coupling relationship between the data in the optimized data according to the indicator dependency relationship; A feature extraction module is used to detect the data time series corresponding to the cluster data, construct a visualization chart of the cluster data based on the data time series, analyze the data trend corresponding to the cluster data according to the visualization chart, and extract the data features of the cluster data, and monitor the data change direction corresponding to the cluster data in real time according to the data trend and the data features; A strategy construction module is used to determine the data requirements corresponding to the cluster data based on the direction of data change, to construct a management strategy corresponding to the cluster data, and to calculate the reliability of the management strategy. When the reliability is greater than a preset threshold, the management strategy corresponding to the reliability is used as the final management strategy for the cluster data.

Images