Multi-dimensional data analysis system and method for data element market

Abstract

The invention discloses a multi-dimensional data analysis system and method for a data element market. The system comprises a data authorization access module, a multi-dimensional feature fusion module, a dynamic pricing analysis module, a data circulation analysis module, a security module and a data valuation module. And the data authorization access module is used for managing a data access interface and an authorization mechanism and allowing standardized access of authenticated multi-dimensional data. The invention belongs to the technical field of data analysis, and aims to solve the problems that value information hidden behind data is difficult to deeply mine and specific dynamic pricing and transaction requirements of a data element market are not fully considered in the prior art. The method has the technical effects that the value information hidden behind the data can be conveniently and deeply mined, meanwhile, the specific dynamic pricing and transaction requirements of the data element market are fully considered, and the dynamic factors of the data element market in real-time change can be conveniently and comprehensively analyzed in time.

Description

Technical Field

[0001] This application belongs to the technical field of data analysis, and specifically relates to a multi-dimensional data analysis system and method for the data element market. Background Art

[0002] In the context of the accelerating digital transformation, data, as a new type of production factor, is as important as traditional factors such as land, labor, and capital. The data element market, as the core carrier for the optimal allocation of data resources, integrates massive data from different industries, fields, and formats to build a multi-level and multi-dimensional data circulation network. This integration not only breaks down the barriers of data silos but also enhances the availability and credibility of data through technical means, laying a foundation for subsequent value mining.

[0003] The efficient circulation and value release of the data element market are gradually becoming important drivers for economic growth and industrial innovation. On the one hand, through in-depth analysis and application of data, enterprises can more accurately understand the needs of the data element market, optimize production processes, and improve decision-making efficiency, thereby promoting the upgrading of traditional industries. On the other hand, the combination of data with other production factors (such as technology, capital, and talent) has given rise to new business forms and models, such as innovative practices in the fields of fintech, intelligent manufacturing, and smart cities. In addition, the development of the data element market has also promoted the improvement of data assetization and trading mechanisms, providing basic support for the quantification of data value and the marketization of data elements.

[0004] However, most current multi-dimensional data analysis systems have obvious limitations in meeting the complex needs of the data element market. They mainly focus on general statistical analysis of data, such as calculating basic indicators like the average, sum, maximum, and minimum of data, making it difficult to deeply explore the value information hidden behind the data. At the same time, they do not fully consider the unique dynamic pricing and trading requirements of the data element market, thus being unable to comprehensively and timely analyze the dynamic factors in real-time change. Summary of the Invention

[0005] This application provides a multi-dimensional data analysis system and method for the data element market, aiming to solve the problems in the prior art that mainly focus on general statistical analysis of data, making it difficult to deeply explore the value information hidden behind the data, and not fully considering the unique dynamic pricing and trading requirements of the data element market, thus being unable to comprehensively and timely analyze the dynamic factors in real-time change.

[0006] In a first aspect, a multi-dimensional data analysis system for the data element market includes a data authorization access module, a multi-dimensional feature fusion module, a dynamic pricing analysis module, a data circulation analysis module, a security module, and a data valuation module;

[0007] The data authorization access module is used to manage data access interfaces and authorization mechanisms, allowing the standardized access of authenticated multi-dimensional data, and preprocessing the accessed multi-dimensional data to obtain an annotation set; the multi-dimensional data includes the dynamic pricing of trading entity suppliers, data demanders, and the database of the data trading platform;

[0008] The multi-dimensional feature fusion module is used to extract the key information of various features in the annotation set, establish fusion rules, and fuse the key information into a unified feature vector;

[0009] The dynamic pricing analysis module can dynamically adjust the pricing of the data factor market based on the dynamic model and the data feature system;

[0010] The data circulation analysis module can classify the multi-faceted data of trading entity suppliers, determine the data type, and calculate the adaptation degree between the data factor and the demander; the data circulation analysis module includes an adaptation sub-module;

[0011] The security module is used to prevent data leakage, tampering, and abuse through security technologies such as encryption and data desensitization;

[0012] The data valuation module is used to make auxiliary trading decisions according to the adaptation degree between the data factor and the demander, and generate a valuation report.

[0013] Further, the data authorization access module includes a multi-dimensional data collection sub-module and a preprocessing sub-module. The multi-dimensional data collection sub-module is used to collect multi-dimensional data from various data sources in the data factor market through the data access interface, and the preprocessing sub-module is used to perform data cleaning, data standardization, and data deduplication on the collected multi-dimensional data.

[0014] Further, the dynamic pricing analysis module includes a dynamic model construction sub-module and an adjustment sub-module. The dynamic model construction sub-module is used to construct a dynamic model and establish a data feature system for the fused key information. The adjustment sub-module is used to adjust the price of the data factor in real time according to the changes in the data factor market and the data quality analysis results output by the dynamic model.

[0015] Further, the adaptation sub-module first needs to set a threshold for the matching degree, then extract the required data type from the demand description of the data demander as the demand feature, classify and code the identified data type, and then calculate the difference between the demander's requirements and the actual quality of the supplier's data factor according to the classification code to obtain the matching degree.

[0016] Further, the data valuation module also includes an interactive interface module, which is used to feedback the generated valuation report to the supplier and the demander respectively.

[0017] Furthermore, the data access interface is a channel for realizing data transmission and interaction between different data sources in the data element market, and is used to define the format, protocol, and method of data transmission.

[0018] Second, a multi-dimensional data analysis method for the data element market includes the following steps:

[0019] S1: Input multi-dimensional data of the data element market through the data access interface, and preprocess the multi-dimensional data to generate an annotation set;

[0020] S2: Extract data quality feature information and data structure feature information from the annotation set, and fuse the data quality feature information and data structure feature information into a unified feature vector based on the defined fusion rules;

[0021] S3: Establish a data feature system according to the unified feature vector, and dynamically adjust the pricing of the data element market based on the dynamic model and the data feature system;

[0022] S4: Determine the data type according to various data of the trading entity supplier, and calculate the matching degree between the data element and the demander;

[0023] S5: Make an auxiliary trading decision according to the matching degree between the data element and the demander and generate a valuation report.

[0024] Furthermore, the specific steps of S1 are as follows:

[0025] S1.1: Define the format, protocol, and method of data transmission of the data access interface;

[0026] S1.2: Data cleaning: Identify missing values in the data, select the median statistical method to fill in the missing values, then detect outliers in the data, and mark the outliers as special cases;

[0027] S1.3: Data standardization: Standardize the cleaned data through Z-score standardization;

[0028] S1.4: Data deduplication: Identify duplicate records in the data by comparing the similarity of the key fields of the data, delete the duplicate records, and then analyze the fields in the data to eliminate the fields irrelevant to the analysis target.

[0029] Furthermore, the specific steps of S4 are as follows;

[0030] S4.1: Based on the clustering algorithm and the elbow method, by plotting the sum of squared errors under different fused key information, select the inflection point as the determined number of clusters K;

[0031] S4.2: Statistically analyze the distribution of characteristics such as the basic information, transaction history data, and credit ratings of the suppliers of the trading entities;

[0032] S4.3: Based on the data types output by the machine learning algorithm sub-module, clarify the degree of adaptation between the data elements and the demand side;

[0033] S4.4: Extract the required data types from the demand descriptions of the data demand side as demand characteristics, classify and code the identified data types, compare the data type codes according to the threshold of the matching degree, and calculate the differences between the requirements of the demand side and the actual quality of the data elements provided by the supply side.

[0034] Compared with the prior art, the present application has at least the following beneficial effects:

[0035] Based on the further analysis and research of the problems in the prior art, the present application collects multi-dimensional data from various data sources in the data element market through the multi-dimensional feature fusion module via the data access interface, focuses on the analysis of the data in the data element market, fully considers the particularity of the data element market, and is used to deeply explore the value information hidden behind the data, providing a solid foundation for subsequent dynamic pricing and transaction analysis. At the same time, the combination of the dynamic pricing analysis module and the data circulation analysis module facilitates a comprehensive and timely analysis of the dynamic factors in the data element market that are in a state of real-time change, which not only helps to capture market dynamics but also promotes the healthy development of the data element market. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] Figure 1 It is a schematic diagram of the modules of a multi-dimensional data analysis system for the data element market provided by an embodiment of the present application;

[0037] Figure 2 It is a schematic flowchart of a multi-dimensional data analysis method for the data element market provided by an embodiment of the present application. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0038] In order to make the objectives, technical solutions, and advantages of the present application clearer and more understandable, the present application will be further described in detail below with reference to the accompanying drawings and embodiments.

[0039] As Figure 1 shown, the multi-dimensional data analysis system for the data element market provided by the present application includes a data authorization access module, a multi-dimensional feature fusion module, a dynamic pricing analysis module, a data circulation analysis module, a security module, and a data valuation module.

[0040] The data authorization access module is used to manage data access interfaces and authorization mechanisms, allowing the standardized access of authenticated multi-dimensional data, and preprocessing the accessed multi-dimensional data to obtain an annotation set, ensuring the legality and compliance of data sources, and providing a reliable basis for subsequent data analysis.

[0041] The data access interface is a channel for realizing data transmission and interaction between different data sources in the data element market, used to define the format, protocol, and method of data transmission, ensuring that multi-dimensional data can be securely, stably, and efficiently transmitted from the source system to this system.

[0042] The authorization mechanism is used to control and manage the access rights of users (suppliers, demanders, and trading platforms) to the data access interface. When a user attempts to access the data access interface, the authorization mechanism will check the user's permissions and decide whether to allow access. If the user does not have the corresponding permissions, the access request will be rejected.

[0043] The data authorization access module includes a multi-dimensional data collection sub-module and a preprocessing sub-module. Among them, the multi-dimensional data collection sub-module is used to collect multi-dimensional data from various data sources in the data element market through the data access interface. The multi-dimensional data includes the dynamic pricing of trading entity suppliers, data demanders, and the databases of data trading platforms.

[0044] Among them, for structured data, such as pricing tables and trading platform data tables, the JDBC interface is used for data extraction; for semi-structured and unstructured data sources, such as web pages and social media platforms, web crawler technology and natural language processing technology are used for data scraping and parsing. Thus, it is convenient to adopt customized collection strategies according to different types of data sources.

[0045] The preprocessing sub-module is used to perform data cleaning, data standardization, and data deduplication on the collected multi-dimensional data, and annotate the processed multi-dimensional data to obtain an annotation set, which can remove irrelevant or redundant information, reduce the data volume, lower the complexity of subsequent processing, and improve the overall operation efficiency of the system. The specific content is as follows:

[0046] a) Data cleaning: Identify missing values in the data, select the median statistical method to fill in the missing values, and then detect outliers in the data, marking the outliers as special cases for attention during subsequent analysis.

[0047] Unify the data types in different data sources into the standard data types required by the system.

[0048] For example, convert numerical strings to numerical types, convert date strings to date types, etc.

[0049] b) Data standardization: The cleaned data is standardized through Z-score standardization, making data of different scales comparable and facilitating subsequent analysis and modeling. Z-score standardization can calculate the Z-score of each data point, that is, the difference between the data point and the mean divided by the standard deviation, converting the data into a distribution with a mean of 0 and a standard deviation of 1. After standardization, data of different scales have the same dimension and range, and can be directly compared and operated on.

[0050] c) Data deduplication: By comparing the similarity of the key fields of the data, duplicate records in the data are identified and deleted. Then, the fields in the data are analyzed to eliminate fields irrelevant to the analysis target, reducing the data volume and the complexity of subsequent processing.

[0051] The multi-dimensional feature fusion module is used to extract the key information of various features in the annotation set, establish fusion rules, and fuse the key information into a unified feature vector to more comprehensively and deeply explore the data value. Among them, the key information includes data quality feature information and data structure feature information, which reflect the basic attributes of the data in the annotation set and provide a basis for subsequent feature combination and analysis. The specific content is as follows:

[0052] Key information: Using the method of graph theory, the fields in the annotation set are used as nodes, and the relationships between the fields are used as edges to construct a field relationship graph. According to the business logic and actual data situation between the fields, the strength of the field relationship is analyzed.

[0053] For example, in the supplier annotation set, there is a primary-foreign key relationship between the customer ID field in the supplier and the customer ID field in the demander, and these two fields can be connected by an edge in the relationship graph.

[0054] Key information fusion: According to the analysis target of the data element market, an association rule between data quality feature information and data structure feature information is established, and the chi-square test statistical method is used to calculate the association strength between data quality features and data structure features.

[0055] For example, fields with a relatively high missing rate are in a non-core position in the data structure and have a weak relationship with some key fields, thus affecting the reliability of the data analysis results related to them.

[0056] Taking the extracted data quality feature information and data structure feature information as feature dimensions, a feature vector is constructed. A feature fusion algorithm is selected to process the feature vector and fuse multiple features into a unified feature vector.

[0057] Among them, feature fusion algorithm generally refers to a type of technical means that integrates multiple heterogeneous features into comprehensive representations to explore deep data associations and improve model performance. It is not limited to a specific algorithm framework or implementation method, and can adapt to a variety of data scenarios and analysis needs, and flexibly use a variety of strategies for fusion. For example, it can be a weighted fusion based on a simple linear combination, which sums up the weights assigned according to the importance of the features to highlight the role of key features. Deep learning networks can also be used for complex nonlinear fusion, and implicit associations between features can be captured through multi-layer neuron interactions.

[0058] The dynamic pricing analysis module can dynamically adjust the pricing of the data factor market based on the dynamic model and data feature system, providing a reasonable reference for subsequent data analysis. Among them, the dynamic pricing analysis module includes a dynamic model construction submodule and an adjustment submodule. The dynamic model construction submodule is used to build a dynamic model and establish a data feature system for the key information of the integration. The specific contents are as follows:

[0059] A dynamic model is built based on a decision tree, making it suitable for processing data with nonlinear relationships and capable of processing categorical and numerical features. The data feature system is constructed based on a unified feature vector value by recursively dividing the fused key information into smaller subsets.

[0060] For example, in data element pricing, the pricing range can be divided according to different combinations of data scarcity (high, medium, low) and data quality (excellent, good, poor).

[0061] The fused key information is divided into a training set and a test set in a ratio of 7-8:2-3. The training set is used to train the dynamic model, and the test set is used to evaluate the performance of the dynamic model.

[0062] The adjustment submodule is used to adjust the price of data elements in real time according to the changes in the data element market and the data quality analysis results output by the dynamic model. The dynamic model trains the adaptive adjustment algorithm through the training set, allowing the algorithm to learn how to adjust the price under different data element market conditions. This enables the dynamic model to adjust the price of data elements in real time based on the real-time monitored data element market data and the decision results of the algorithm. For example, when the data element market demand heat index suddenly rises above the threshold, the algorithm determines the extent of the price increase based on the learned strategy.

[0063] The data circulation analysis module can classify various aspects of the transaction subject supplier data, determine the data type, calculate the degree of adaptation of the data elements and the demand side, and provide a basis for subsequent analysis. The data circulation analysis module includes a machine learning algorithm submodule and an adaptation submodule.

[0064] The machine learning algorithm sub-module uses a clustering algorithm to classify the multi-faceted data of the trading entity supplier in the fused key information, and describes different types of trading entity suppliers. The specific content is as follows:

[0065] a) Determine the number of clusters K

[0066] Use the elbow method. By plotting the sum of squared errors (SSE) under different fused key information, select the inflection point as the determined number of clusters K. The calculation formula for the sum of squared errors (SSE) is as follows:

[0067]

[0068] where i represents the i-th cluster, p is the sample point, and is the mean of the samples in the i-th cluster.

[0069] The lower the SSE value, the better the clustering effect. Plot a curve with the k value as the abscissa and SSE as the ordinate. As k increases, SSE will gradually decrease, but the rate of decrease will gradually slow down. At the inflection point of the curve (i.e., the point where the SSE decline rate significantly slows down), the corresponding k value is the optimal number of clusters.

[0070] b) Determine the types of trading entity suppliers

[0071] For the trading entity suppliers within each cluster, count the distribution of their basic information, trading history data, credit ratings, and other characteristics. Based on the significant characteristics of each cluster, name the clustering clusters to more intuitively understand the types of trading entity suppliers. Check whether the trading entity suppliers within the same cluster have similar characteristics and behavior patterns, and whether the differences between different clusters are obvious. The higher the within-cluster similarity and the lower the between-cluster similarity, the better the clustering effect and the classification effect.

[0072] The adaptation sub-module is used to clarify the adaptation degree between the data elements and the demand side according to the data type output by the machine learning algorithm sub-module. The specific content is as follows:

[0073] First, set the threshold of the matching degree. Only trading pairs with a matching degree greater than the threshold are considered potential matching transactions, and then it is judged whether the trading entity supplier meets the requirements of the demand side.

[0074] Then, extract the required data types from the demand description of the data demand side as demand characteristics, and classify and code the identified data types.

[0075] For example, structured data is encoded as 1, unstructured data is encoded as 2, and semi-structured data is encoded as 3. If the requester needs multiple data types, multi-value encoding or vector encoding can be used. For example, if both structured and unstructured data are needed, it can be encoded as [1, 2, 0]. If the data type encodings provided by the requester and the supplier are the same, the similarity is 1; if they are different, the similarity is 0.

[0076] Calculate the difference between the requirements of the requester and the actual quality of the data elements provided by the supplier according to the classification encoding to obtain the matching degree.

[0077] For example, if the requester requires an accuracy of 0.9 and the accuracy of the data elements provided by the supplier is 0.85, the similarity of this indicator is 1 - ∣0.9 - 0.85∣ = 0.95. The similarities of each quality indicator are weighted and averaged to obtain the matching degree of data quality, which is convenient for judging whether the trading entity supplier meets the requirements of the requester according to the matching degree.

[0078] The security module is used to prevent data leakage, tampering and abuse through security technologies such as encryption and data desensitization, ensure the security of data during storage, transmission and use, safeguard the rights and interests of data owners and users, and at the same time establish a strict access control mechanism. Different access permissions are assigned to users according to their roles and responsibilities. The permission management adopts a fine-grained access control strategy to precisely control users' access permissions to different data resources and analysis functions. For example, the trading entity supplier can only access and modify the data element information it provides, and the data requester can only view the data and analysis results related to its own transactions.

[0079] The data valuation module is used to make auxiliary trading decisions according to the adaptability of data elements to the requester, and can provide a basis for the trading entity supplier and a reference for the requester's procurement respectively. Provide data pricing suggestions for the supplier, for example: Suggested quotation range: Based on the valuation result and the market average price, recommend a reasonable quotation range and generate a valuation report. Provide procurement reference for the requester, for example: Cost performance ranking: Based on the valuation result and the market average price, recommend a list of data elements that meet the conditions and generate a valuation report:

[0080] At the same time, the data valuation module also includes an interactive interface module, which is used to feedback the generated valuation reports to the supplier and the requester respectively, and the specific content is as follows:

[0081] Supplier perspective interface: Provide a dedicated interactive interface for the supplier to display the valuation results of the data elements it provides in different demand scenarios. The interface includes the basic information of the data elements and the content of the matching degree analysis. The supplier can adjust the quality, price or promotion strategy of the data elements according to the analysis content.

[0082] Demand-side perspective interface: A personalized interface is designed for the demand side to display a list of eligible data elements based on their business needs and procurement preferences. The interface provides detailed data element information, valuation comparison, supplier evaluation, and simulated procurement functions. The demand side can filter, sort, and compare different data elements through the interface, perform simulated procurement operations, and view the costs, benefits, and risks under different procurement plans.

[0083] Multi-role collaborative interaction: Supports collaborative interaction between suppliers and demanders on the same platform. For example, both parties can communicate in real time, negotiate prices and transaction terms, share data samples, etc. through the interface. At the same time, the platform can record the interaction history between the two parties to provide reference for subsequent transaction decisions.

[0084] In the above-mentioned multidimensional data analysis system for the data element market, the multidimensional feature fusion module collects multidimensional data from various data sources in the data element market through the data access interface, focusing on the analysis of data in the data element market, taking full account of the particularity of the data element market, and is used to deeply explore the value information hidden behind the data, providing a solid foundation for subsequent dynamic pricing and transaction analysis. At the same time, the dynamic pricing analysis module is combined with the data circulation analysis module to facilitate a comprehensive and timely analysis of the dynamic factors in the data element market that are in real-time change, which not only helps to capture market dynamics, but also promotes the healthy development of the data element market.

[0085] like Figure 2 As shown, the multidimensional data analysis method for the data element market provided by this application specifically includes the following steps:

[0086] S1: Input the multidimensional data of the data factor market through the data access interface, pre-process the multidimensional data, and generate a labeling set. At the same time, when inputting multidimensional data, an authentication mechanism is required to ensure that the data source is legal and compliant.

[0087] The specific steps of S1 are as follows:

[0088] S1.1: Define the format, protocol and method of data transmission of the data access interface, allowing standardized access to authenticated multi-dimensional data;

[0089] S1.2: Data cleaning: Identify missing values ​​in the data, select the median statistical method to fill in the missing values, then detect outliers in the data, mark the outliers as special cases, so that they can be noted in subsequent analysis. Convert the data types in different data sources to the standard data types required by the system.

[0090] S1.3: Data Standardization: The cleaned data is standardized through Z-score standardization to make data of different scales comparable. Z-score standardization can calculate the Z-score of each data point, that is, the difference between the data point and the mean divided by the standard deviation, and convert the data into a distribution with a mean of 0 and a standard deviation of 1. After standardization, data of different scales have the same dimension and range, and can be directly compared and operated on.

[0091] S1.4: Data Duplication Removal: By comparing the similarity of the key fields of the data, duplicate records in the data are identified and deleted. Then, the fields in the data are analyzed, and the fields irrelevant to the analysis target are removed to reduce the data volume and the complexity of subsequent processing.

[0092] S2: Extract data quality feature information and data structure feature information from the annotation set, and fuse the data quality feature information and the data structure feature information into a unified feature vector based on the defined fusion rules;

[0093] S3: Establish a data feature system according to the unified feature vector, and dynamically adjust the pricing of the data factor market based on the dynamic model and the data feature system to provide a reasonable reference for subsequent data analysis. The dynamic model can adjust the price of data factors in real time according to the real-time monitored data of the data factor market and the decision results of the algorithm.

[0094] S4: Determine the data type according to various data of the trading entity supplier, and calculate the matching degree between the data factor and the demander.

[0095] The specific steps of S4 are as follows:

[0096] S4.1: Based on the clustering algorithm and the Elbow Method, by plotting the sum of squared errors (SSE) under different fused key information, select the number of clusters K determined by the inflection point;

[0097] S4.2: Statistically analyze the distribution of characteristics such as the basic information, trading history data, and credit rating of the trading entity supplier;

[0098] S4.3: According to the data type output by the machine learning algorithm sub-module, clarify the matching degree between the data factor and the demander;

[0099] S4.4: Extract the required data type from the demand description of the data demander as the demand feature, classify and code the identified data types, compare the data type codes according to the threshold of the matching degree, and calculate the difference between the requirements of the demander and the actual quality of the data factors provided by the supplier.

[0100] S5: Make an auxiliary transaction decision and generate a valuation report according to the degree of adaptation between the data element and the demander, provide a quotation suggestion for the supplier, and provide a cost performance ranking for the demander. At the same time, be able to feedback the generated valuation report to the supplier and the demander respectively.

[0101] The technical features of the above embodiments can be combined arbitrarily. For the sake of concise description, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, it should be considered as the scope recorded in this specification.

Claims

1. A multi-dimensional data analysis system for the data element market, characterized in that Including: A data authorization access module, a multi-dimensional feature fusion module, a dynamic pricing analysis module, a data circulation analysis module, a security module, and a data valuation module; The data authorization access module is used to manage data access interfaces and authorization mechanisms, allowing the standardized access of authenticated multi-dimensional data, and preprocessing the accessed multi-dimensional data to obtain an annotation set; the multi-dimensional data includes the dynamic pricing of trading entity suppliers, data demanders, and the database of the data trading platform; The multi-dimensional feature fusion module is used to extract the key information of various features in the annotation set, establish a fusion rule, and fuse the key information into a unified feature vector; The dynamic pricing analysis module can dynamically adjust the pricing of the data factor market based on the dynamic model and the data feature system; The data circulation analysis module can classify the multi-faceted data of trading entity suppliers, determine the data type, and calculate the matching degree between the data factor and the demander; the data circulation analysis module includes an adaptation sub-module; The security module is used to prevent data leakage, tampering, and abuse through security technologies such as encryption and data desensitization; The data valuation module is used to make auxiliary trading decisions according to the matching degree between the data factor and the demander, and generate a valuation report.

2. The multi-dimensional data analysis system for the data element market according to claim 1, wherein The data authorization access module includes a multi-dimensional data collection sub-module and a preprocessing sub-module. The multi-dimensional data collection sub-module is used to collect multi-dimensional data from various data sources in the data factor market through the data access interface, and the preprocessing sub-module is used to perform data cleaning, data standardization, and data deduplication on the collected multi-dimensional data.

3. The multi-dimensional data analysis system for the data element market according to claim 1, wherein The dynamic pricing analysis module includes a dynamic model construction sub-module and an adjustment sub-module. The dynamic model construction sub-module is used to construct a dynamic model and establish a data feature system for the fused key information, and the adjustment sub-module is used to adjust the price of the data factor in real time according to the changes in the data factor market and the data quality analysis results output by the dynamic model.

4. The multi-dimensional data analysis system for the data element market according to claim 1, characterized in that, The adaptation sub-module first needs to set a threshold for the matching degree, then extract the required data type from the demand description of the data demander as a demand feature, classify and code the identified data type, and then calculate the difference between the demander's requirements and the actual quality of the supplier's data factor according to the classification code to obtain the matching degree.

5. The multi-dimensional data analysis system for the data element market according to claim 1, characterized in that The data valuation module further includes an interactive interface module, and the interactive interface module is used to feedback the generated valuation report to the supplier and the demander respectively.

6. The multi-dimensional data analysis system for the data element market according to claim 1, characterized in that The data access interface is a channel for realizing data transmission and interaction between different data sources in the data factor market, and is used to define the format, protocol, and method of data transmission.

7. Multidimensional data analysis method for the data element market, characterized in that Including the following steps: S1: Input the multi-dimensional data of the data factor market through the data access interface, and preprocess the multi-dimensional data to generate an annotation set; S2: Extract the data quality feature information and data structure feature information from the annotation set, and fuse the data quality feature information and data structure feature information into a unified feature vector based on the defined fusion rule; S3: Establish a data feature system according to the unified feature vector, and dynamically adjust the pricing of the data factor market based on the dynamic model and the data feature system; S4: Determine the data type based on various data of the trading entity supplier, and calculate the degree of fit between the data element and the demander. S5: Make an auxiliary trading decision and generate a valuation report according to the degree of fit between the data element and the demander.

8. The multi-dimensional data analysis method for the data element market according to claim 7, wherein The specific steps of S1 are as follows: S1.1: Define the data transmission format, protocol and method of the data access interface. S1.2: Data cleaning: Identify the missing values in the data, select the median statistical method to fill the missing values, then detect the outliers in the data, and mark the outliers as special cases. S1.3: Data standardization: Standardize the cleaned data through Z-score standardization. S1.4: Data deduplication: Identify the duplicate records in the data by comparing the similarity of the key fields of the data, delete the duplicate records, and then analyze the fields in the data to eliminate the fields irrelevant to the analysis target.

9. The multi-dimensional data analysis method for the data element market according to claim 7, characterized in that, The specific steps of S4 are as follows; S4.1: Based on the clustering algorithm and the elbow method, by plotting the sum of squared errors under different fusions of key information, select the inflection point as the determined number of clusters K. S4.2: Statistically analyze the distribution of characteristics such as the basic information, trading history data and credit rating of the trading entity supplier. S4.3: Determine the degree of fit between the data element and the demander according to the data type output by the machine learning algorithm sub-module. S4.4: Extract the required data type from the demand description of the data demander as the demand feature, classify and code the identified data type, compare the data type codes according to the threshold of the matching degree, and calculate the difference between the requirements of the demander and the actual quality of the data elements of the supplier.

Images