A big data analysis and decision-making system based on data elements

Abstract

A big data analysis and decision-making system based on data elements, relating to the technical field of data analysis; element processing is performed on various original data of an enterprise to obtain different data element categories, obtain various historical decision information of the enterprise and its decision return ratio, obtain the data element categories and data element proportions contained in various historical decision information, construct a decision analysis diagram and obtain the gain coefficient of each data element category for the decision return ratio, obtain current decision information and input it into a preset return evaluation model to obtain a predicted return ratio, perform element optimization on the data element categories and data element proportions in the current decision information and obtain an optimized return ratio; it is beneficial to represent the potential gain situation through quantitative values, and can provide a reference for data elements for relevant personnel when making decisions.

Description

Technical Field

[0001] The present invention relates to the field of data analysis technology, and specifically to a big data analysis and decision-making system based on data elements. Background Art

[0002] As a hub for data circulation, the digital platform uses the original data of enterprises as a carrier, connects with government and industrial park data, obtains targeted and open data resources, and provides computing power + data + algorithms to provide support services for various application scenarios. It also conducts industry analysis and research, industrial decision-making analysis, and enterprise consulting and other related services through offline think tanks to solve problems such as insufficient investment in digital construction factors, imperfect industrial chains, and limited industrial park development in the development process of key industries.

[0003] In the prior art, it has become a mature technology to use data elements through digital platforms to provide data references for enterprises' decision-making behaviors. However, the prior art often finds it difficult to obtain the degree of gain of various data elements on the park's revenue, and therefore cannot provide accurate and targeted decision-making assistance, that is, it is impossible to provide decision-making assistance at the specific data element level. In view of the shortcomings of the prior art, the present invention provides a big data analysis and decision-making system based on data elements. Summary of the invention

[0004] The purpose of the present invention is to provide a big data analysis and decision-making system based on data elements.

[0005] The purpose of the present invention can be achieved through the following technical solution: A big data analysis and decision-making system based on data elements, comprising the following modules:

[0006] The data collection module is used to obtain various original data of the enterprise and upload them to the digital platform, and perform element processing on the various original data to obtain different data element categories;

[0007] The decision acquisition module is used to obtain the enterprise's historical decision information and its corresponding decision return ratio, and obtain the data element categories and data element proportions contained in each historical decision information;

[0008] A decision analysis module is used to construct a corresponding decision analysis diagram based on the decision return ratio and data element proportion of each historical decision information, and obtain the gain coefficient of each data element category for the decision return ratio in the decision analysis diagram;

[0009] The data association module is used to obtain the current decision information and its data element categories and data element proportions, and input them into the preset return evaluation model to obtain the corresponding predicted return ratio;

[0010] The decision optimization module is used to optimize the data element categories and data element proportions in the current decision information according to the predicted return ratio, and obtain the corresponding optimized return ratio.

[0011] Furthermore, the process of obtaining various original data of the enterprise and uploading them to the digital platform includes:

[0012] A collection unit is set up to collect various original data of the enterprise respectively, wherein the original data comes from public data, service enterprise data, and public data of cooperative government parks, so as to build a digital platform for the enterprise, and upload the collected original data to the digital platform for storage.

[0013] Furthermore, the process of factorizing each raw data to obtain different data element categories includes:

[0014] The element processing refers to performing multiple data processing on each item of original data to classify it into different data element categories, and the multiple data processing includes data cleaning, data conversion, and data visualization on the original data;

[0015] Different data features are extracted from the original data, including structural features, source features, value features, and scenario features. The original data is divided into different data element categories based on the extracted data features. The data element categories are used to reflect the data features contained in different original data.

[0016] Furthermore, the process of obtaining various historical decision information of the enterprise and its corresponding decision return ratio, and obtaining the data element categories and data element proportions contained in various historical decision information includes:

[0017] The historical decision information refers to the sum of decision processes that have occurred and are recorded and their related raw data. The investment income and investment cost of the decision process corresponding to a single historical decision information are obtained, and the decision return ratio of the historical decision information is obtained based on the investment income and investment cost.

[0018] The data element categories corresponding to each item of original data in a single historical decision information are collectively taken as the data element categories included in the historical decision information, and the data element ratio of each data element category included in the historical decision information is obtained;

[0019] The data element proportion is the ratio of the total number of a single data element category in the historical decision information in each item of original data to the total number of all data element categories in each item of original data.

[0020] Furthermore, the process of constructing a corresponding decision analysis diagram based on the decision return ratio and data element proportion of each historical decision information includes:

[0021] A regular polygon is constructed according to the number of data element categories in the historical decision information, wherein the distance between each vertex and the center of the regular polygon is a preset fixed distance, and a three-dimensional coordinate system of the historical decision information is constructed with the center of the regular polygon as the coordinate origin, wherein the regular polygon falls on a plane formed by the x-axis and the y-axis;

[0022] The center of the regular polygon corresponds to the decision return ratio of the historical decision information, and the coordinate point corresponding to the value of the decision return ratio is reflected on the z-axis. Each vertex of the regular polygon corresponds to a data element category, and the coordinate point corresponding to the product of the proportion of data elements in each data element category and the investment income of the historical decision information is reflected in the three-dimensional coordinate system. The three-dimensional coordinate system at this time is used as the decision analysis diagram of the historical decision information.

[0023] Furthermore, the process of obtaining the gain coefficient of each data element category for the decision return ratio in the decision analysis diagram includes:

[0024] Obtain decision analysis diagrams for different historical decision information, and obtain the gain sub-coefficient B of the decision-return ratio of a single data element category in each decision analysis diagram i , where i = 1, 2, …, n, and n represents the number of decision analysis diagrams containing this single data element category;

[0025]

[0026] Among them, M i represents the number of all data element categories in the i-th decision analysis diagram, L i Represents the shortest distance between the coordinate point of the single data element category in the i-th decision analysis diagram and the coordinate point of its decision return ratio, G i Represents the minimum area enclosed by the coordinate point of the single data element category in the i-th decision analysis diagram and the x-axis, y-axis, and z-axis. are preset fixed parameters;

[0027] According to the gain sub-coefficients of the single data element category in each decision analysis diagram, the gain coefficient D of the single data element category for the decision return ratio is obtained;

[0028]

[0029] The gain sub-coefficients of each data element category in the decision analysis diagram containing it are obtained respectively, and the gain coefficients of each data element category for the decision return ratio are obtained according to the gain sub-coefficients.

[0030] Furthermore, the process of obtaining the current decision information and its data element categories and data element proportions, and inputting them into a preset return evaluation model to obtain the corresponding predicted return ratio includes:

[0031] The current decision information refers to the sum of the ongoing decision process and its related raw data, and the categories of data elements and their proportions contained in the current decision information are obtained;

[0032] The preset process of the reward evaluation model is as follows: generating a reward evaluation set according to the data element categories and their data element proportions and gain coefficients in different decision analysis diagrams, and the corresponding decision-return ratios, and dividing the reward evaluation set into a training set and a test set;

[0033] Construct a convolutional neural network, use the data element categories and their data element proportions and gain coefficients in different decision analysis diagrams in the training set as input data of the convolutional neural network, use the corresponding decision return ratios in the training set as output data of the convolutional neural network, and train the convolutional neural network to obtain an initial convolutional neural network;

[0034] The initial convolutional neural network is model verified using the test set, and the initial convolutional neural network with an output value less than or equal to a preset test error threshold is used as the reward evaluation model.

[0035] Furthermore, the process of optimizing the data element categories and data element proportions in the current decision information according to the predicted return ratio and obtaining the corresponding optimized return ratio includes:

[0036] Set an expected return ratio standard for the current decision information. When the predicted return ratio is less than or equal to the expected return ratio standard, optimize the data element categories and data element proportions in the current decision information.

[0037] The factor optimization refers to adding data factor categories and increasing the proportion of data factors on the basis of current decision-making information, and generating different optimization plans, using the return evaluation model to obtain the predicted return ratio under each optimization plan, and taking the largest predicted return ratio that is greater than the expected return ratio standard as the optimized return ratio, and feeding back the optimization plan corresponding to the optimized return ratio to relevant personnel.

[0038] Compared with the prior art, the present invention has the following beneficial effects:

[0039] 1. The present invention processes the original data of the enterprise into elements to divide it into different data element categories, and obtains historical decision information and its decision return ratio at the same time, so as to associate the benefits brought by different decisions with the data element categories. By constructing a decision analysis diagram, the gain coefficients of different data element categories and their data element proportions for the decision return ratio can be obtained, which is conducive to expressing the potential gain situation through quantitative values;

[0040] 2. The present invention constructs a return evaluation model, and uses the return evaluation model to evaluate the predicted return ratio of the current decision information according to the data element categories, data element proportions and gain coefficients in the current decision information, and optimizes the data element categories and data element proportions in the current decision information according to the predicted return ratio, which is conducive to obtaining the maximum optimized return ratio and can provide relevant personnel with a reference for data elements when making decisions. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] Figure 1 It is a schematic diagram of the present invention. DETAILED DESCRIPTION

[0042] like Figure 1 As shown, a big data analysis and decision-making system based on data elements includes the following modules:

[0043] The data collection module is used to obtain various original data of the enterprise and upload them to the digital platform, and perform element processing on the various original data to obtain different data element categories;

[0044] The decision acquisition module is used to obtain the enterprise's historical decision information and its corresponding decision return ratio, and obtain the data element categories and data element proportions contained in each historical decision information;

[0045] A decision analysis module is used to construct a corresponding decision analysis diagram based on the decision return ratio and data element proportion of each historical decision information, and obtain the gain coefficient of each data element category for the decision return ratio in the decision analysis diagram;

[0046] The data association module is used to obtain the current decision information and its data element categories and data element proportions, and input them into the preset return evaluation model to obtain the corresponding predicted return ratio;

[0047] The decision optimization module is used to optimize the data element categories and data element proportions in the current decision information according to the predicted return ratio, and obtain the corresponding optimized return ratio.

[0048] It should be further explained that, in the specific implementation process, the process of obtaining the original data of the enterprise and uploading it to the digital platform includes:

[0049] A collection unit is set up, and various original data of the enterprise are collected through the collection unit. The original data comes from public data, service enterprise data, and public data of cooperative government parks. A digital platform for the enterprise is built, and the collected original data are uploaded to the digital platform for storage. The data has not yet been processed in subsequent data processing, so it is regarded as original data;

[0050] The original data includes policy information, basic enterprise information, enterprise land area information, enterprise production and operation information, enterprise scale structure information, enterprise scientific and technological innovation information, enterprise human resources and social security information, enterprise open cooperation information, enterprise energy conservation and environmental protection information, enterprise production safety information, park land information, etc.

[0051] It should be further explained that, in the specific implementation process, the process of factorizing each raw data to obtain different data element categories includes:

[0052] The element processing refers to performing multiple data processing on each item of original data to classify it into different data element categories, and the multiple data processing includes data cleaning, data conversion, and data visualization on the original data;

[0053] The data cleaning is used to process missing values, duplicate values, and outliers to ensure the accuracy and consistency of the original data, and to merge the original data from different sources into a unified data set;

[0054] The data conversion is used to convert the original data into a format suitable for analysis, and perform descriptive analysis, diagnostic analysis, and normative analysis on the original data, thereby extracting different data features from the original data, including structural features, source features, value features, and scenario features;

[0055] The data visualization refers to the use of data visualization tools to display raw data in the form of charts, graphs, knowledge graphs, etc., so as to facilitate understanding and analysis by relevant personnel;

[0056] The original data is divided into different data element categories according to the extracted data features. Specifically, the original data can be divided into structured data, semi-structured data, and unstructured data according to the structural features, and can be divided into internal enterprise data, external market data, and IoT data according to the source features;

[0057] The raw data can be divided into high-value data (such as purchasing behavior and preferences, sales and sales channels), medium-value data (such as production plans and production progress), and low-value data (such as temporary files and log data) according to value characteristics. The specific decision is made by relevant personnel;

[0058] According to the scene characteristics, the original data can be divided into park data (including park statistical data, park collection data, park commercial supporting data), industry data (including industry digital maps, industry association networks, industry structure maps, industry development indexes), and government data (including public security data, financial management data, land and environmental protection data). The data element categories are used to reflect the data characteristics contained in different original data.

[0059] It should be further explained that, in the specific implementation process, the process of obtaining the enterprise's historical decision information and its corresponding decision-return ratio, and obtaining the data element categories and data element proportions contained in each historical decision information includes:

[0060] The historical decision information refers to the sum of decision-making processes that have occurred and are recorded and their related raw data, which includes different raw data;

[0061] Obtain the investment income and investment cost of the decision-making process corresponding to a single historical decision information, and obtain the decision return ratio corresponding to the historical decision information based on the investment income and investment cost, denoted as R;

[0062]

[0063] Since a single historical decision information includes different original data, the data element categories corresponding to each of the original data can be collectively used as the data element categories included in the historical decision information;

[0064] For example, if the historical decision information includes original data a, original data b, and original data c, and original data a can be classified into data element categories s1, s2, and s3, original data b can be classified into data element categories s2, s3, s5, and s6, and original data c can be classified into data element categories s2 and s5, then the historical decision information includes data element categories s1, s2, s3, s4, s5, and s6;

[0065] Obtaining the data element ratio of each data element category contained in the historical decision information, where the data element ratio is the ratio of the total number of a single data element category in each original data item to the total number of all data element categories in each original data item in the historical decision information;

[0066] For example, the total number of data element category s1 in each item of original data is 1, the total number of data element category s2 in each item of original data is 3, the total number of data element category s3 in each item of original data is 2, and the total number of all data element categories in each item of original data is 9. Then, the proportion of data elements in the historical decision information by data element category s1 is 1 / 9, the proportion of data elements in data element category s2 is 3 / 9, and the proportion of data elements in data element category s3 is 2 / 9.

[0067] It should be further explained that, in the specific implementation process, the process of constructing the corresponding decision analysis diagram according to the decision return ratio and data element proportion of each historical decision information includes:

[0068] Taking any historical decision information as an example, obtain the decision return ratio, data element category, and data element proportion corresponding to the historical decision information, and construct a regular polygon according to the number of its data element categories, wherein the number of sides of the constructed regular polygon is the same as the number of data element categories;

[0069] The distance between each vertex and the center of the regular polygon is a preset fixed distance, and a three-dimensional coordinate system of the historical decision information is constructed with the center of the regular polygon as the coordinate origin. The regular polygon falls on the plane formed by the x-axis and the y-axis. A decision analysis diagram of the historical decision information is constructed on the basis of the three-dimensional coordinate system. The specific steps are as follows:

[0070] The center of the regular polygon corresponds to the decision return ratio of the historical decision information, and the coordinate point corresponding to the value of the decision return ratio is reflected on the z-axis. Each vertex of the regular polygon corresponds to a data element category, and the coordinate point corresponding to the product of the proportion of data elements in each data element category and the investment income of the historical decision information is reflected in the three-dimensional coordinate system. The three-dimensional coordinate system at this time is used as the decision analysis diagram of the historical decision information.

[0071] It should be further explained that, in the specific implementation process, the process of obtaining the gain coefficient of each data element category for the decision return ratio in the decision analysis diagram includes:

[0072] Obtain decision analysis diagrams of different historical decision information, and obtain the gain sub-coefficient of the decision-return ratio of a single data element category in each decision analysis diagram, denoted as B i , where i = 1, 2, …, n, and n represents the number of decision analysis diagrams containing this single data element category;

[0073]

[0074] Among them, M irepresents the number of all data element categories in the i-th decision analysis diagram, that is, the number of sides of its regular polygon, L i Represents the shortest distance between the coordinate point of the single data element category in the i-th decision analysis diagram and the coordinate point of its decision return ratio, G i Represents the minimum area enclosed by the coordinate point of the single data element category in the i-th decision analysis diagram and the x-axis, y-axis, and z-axis;

[0075] in, All are preset fixed parameters. According to the gain sub-coefficient of the single data element category in each decision analysis diagram, the gain coefficient of the single data element category for the decision return ratio is obtained, which is recorded as D;

[0076]

[0077] This method is adopted to obtain the gain sub-coefficient of each data element category in the decision analysis diagram containing it, and then obtain the gain coefficient of each data element category for the decision return ratio.

[0078] It should be further explained that, in the specific implementation process, the process of obtaining the current decision information and its data element categories and data element proportions, and inputting them into the preset return evaluation model to obtain the corresponding predicted return ratio includes:

[0079] The current decision information refers to the sum of the ongoing decision process and its related raw data, which also includes different raw data, and obtains the data element categories and data element proportions included in the current decision information;

[0080] The preset process of the reward evaluation model is as follows: generating a reward evaluation set according to the data element categories and their data element proportions and gain coefficients in different decision analysis diagrams, and the corresponding decision-return ratios, and dividing the obtained reward evaluation set into a training set and a test set;

[0081] Construct a convolutional neural network, use the data element categories and their data element proportions and gain coefficients in different decision analysis diagrams in the training set as input data of the convolutional neural network, use the corresponding decision return ratios in the training set as output data of the convolutional neural network, and train the convolutional neural network to obtain an initial convolutional neural network;

[0082] The initial convolutional neural network is verified using the test set, and the initial convolutional neural network with a value less than or equal to the preset test error threshold is output as the corresponding reward evaluation model. The data element categories, their data element proportions and gain coefficients in the current decision information are input into the reward evaluation model, and the corresponding predicted return ratio is output through the reward evaluation model.

[0083] It should be further explained that, in the specific implementation process, the process of optimizing the data element categories and data element proportions in the current decision information according to the predicted return ratio and obtaining the corresponding optimized return ratio includes:

[0084] Set an expected return ratio standard for the current decision information, compare the obtained predicted return ratio with the expected return ratio standard, if the predicted return ratio is greater than the expected return ratio standard, do not perform any other operations on it, if the predicted return ratio is less than or equal to the expected return ratio standard, optimize the data element categories and data element proportions in the current decision information;

[0085] The element optimization refers to increasing the data element categories and improving the proportion of data elements on the basis of current decision-making information, and generating different optimization schemes. The increase in data element categories and the increase in data element proportions under different optimization schemes are different;

[0086] The return evaluation model is used to obtain the predicted return ratio under each optimization plan, and the predicted return ratio that is greater than the expected return ratio standard and the largest is marked as the optimized return ratio. The optimization plan corresponding to the optimized return ratio is fed back to relevant personnel for reference.

[0087] The above embodiments are only used to illustrate the technical method 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 method of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical method of the present invention.

Claims

1. A big data analysis and decision-making system based on data elements, characterized in that: Includes the following modules: The data collection module is used to obtain various original data of the enterprise and upload them to the digital platform, and perform element processing on the various original data to obtain different data element categories; The decision acquisition module is used to obtain the enterprise's historical decision information and its corresponding decision return ratio, and obtain the data element categories and data element proportions contained in each historical decision information; A decision analysis module is used to construct a corresponding decision analysis diagram based on the decision return ratio and data element proportion of each historical decision information, and obtain the gain coefficient of each data element category for the decision return ratio in the decision analysis diagram; The data association module is used to obtain the current decision information and its data element categories and data element proportions, and input them into the preset return evaluation model to obtain the corresponding predicted return ratio; The decision optimization module is used to optimize the data element categories and data element proportions in the current decision information according to the predicted return ratio, and obtain the corresponding optimized return ratio; The process of obtaining historical decision information and its decision return ratio, data element categories, and data element proportions includes: The historical decision information refers to the sum of decision processes that have occurred and are recorded and their related raw data. The investment income and investment cost of the decision process corresponding to a single historical decision information are obtained, and the decision return ratio of the historical decision information is obtained based on the investment income and investment cost. The data element categories corresponding to each item of original data in a single historical decision information are collectively taken as the data element categories included in the historical decision information, and the data element ratio of each data element category included in the historical decision information is obtained; The data element ratio is the ratio of the total number of a single data element category in the historical decision information in each item of original data to the total number of all data element categories in each item of original data; The process of constructing a corresponding decision analysis diagram based on historical decision information includes: A regular polygon is constructed according to the number of data element categories in the historical decision information, wherein the distance between each vertex and the center of the regular polygon is a preset fixed distance, and a three-dimensional coordinate system of the historical decision information is constructed with the center of the regular polygon as the coordinate origin, wherein the regular polygon falls on a plane formed by the x-axis and the y-axis; The center of the regular polygon corresponds to the decision return ratio of the historical decision information, and the coordinate point corresponding to the value of the decision return ratio is reflected on the z-axis. Each vertex of the regular polygon corresponds to a data element category, and the coordinate point corresponding to the product of the data element ratio of each data element category and the investment income of the historical decision information is reflected in the three-dimensional coordinate system. The three-dimensional coordinate system at this time is used as the decision analysis diagram of the historical decision information; The process of obtaining the gain coefficient of each data element category for the decision-making return ratio includes: Obtain decision analysis diagrams for different historical decision information, and obtain the gain sub-coefficient B of the decision-return ratio of a single data element category in each decision analysis diagram i , where i = 1, 2, …, n, and n represents the number of decision analysis diagrams containing this single data element category; Among them, M i represents the number of all data element categories in the i-th decision analysis diagram, L i Represents the shortest distance between the coordinate point of the single data element category in the i-th decision analysis diagram and the coordinate point of its decision return ratio, G i Represents the minimum area enclosed by the coordinate point of the single data element category in the i-th decision analysis diagram and the x-axis, y-axis, and z-axis. are preset fixed parameters; According to the gain sub-coefficients of the single data element category in each decision analysis diagram, the gain coefficient D of the single data element category for the decision return ratio is obtained; Obtain the gain sub-coefficient of each data element category in the decision analysis diagram containing it, and obtain the gain coefficient of each data element category for the decision return ratio according to the gain sub-coefficient; The process of obtaining current decision information and inputting it into the return evaluation model to obtain the predicted return ratio includes: The current decision information refers to the sum of the ongoing decision process and its related raw data, and the categories of data elements and their proportions contained in the current decision information are obtained; The preset process of the reward evaluation model is as follows: generating a reward evaluation set according to the data element categories and their data element proportions and gain coefficients in different decision analysis diagrams, and the corresponding decision-return ratios, and dividing the reward evaluation set into a training set and a test set; Construct a convolutional neural network, use the data element categories and their data element proportions and gain coefficients in different decision analysis diagrams in the training set as input data of the convolutional neural network, use the corresponding decision return ratios in the training set as output data of the convolutional neural network, and train the convolutional neural network to obtain an initial convolutional neural network; The initial convolutional neural network is model verified using the test set, and the initial convolutional neural network with an output value less than or equal to a preset test error threshold is used as the reward evaluation model.

2. The big data analysis and decision-making system based on data elements according to claim 1, characterized in that: The process of obtaining raw data and uploading it to a digitization platform includes: A collection unit is set up to collect various original data of the enterprise respectively, wherein the original data comes from public data, service enterprise data, and public data of cooperative government parks, so as to build a digital platform for the enterprise, and upload the collected original data to the digital platform for storage.

3. The big data analysis and decision-making system based on data elements according to claim 2 is characterized in that: The process of factorizing the original data to obtain data element categories includes: The element processing refers to performing multiple data processing on each item of original data to classify it into different data element categories, and the multiple data processing includes data cleaning, data conversion, and data visualization on the original data; Different data features are extracted from the original data, including structural features, source features, value features, and scenario features. The original data is divided into different data element categories based on the extracted data features. The data element categories are used to reflect the data features contained in different original data.

4. The big data analysis and decision-making system based on data elements according to claim 3 is characterized in that: The process of optimizing the elements of current decision information and obtaining the optimal return ratio includes: Set an expected return ratio standard for the current decision information. When the predicted return ratio is less than or equal to the expected return ratio standard, optimize the data element categories and data element proportions in the current decision information. The factor optimization refers to adding data factor categories and increasing the proportion of data factors on the basis of current decision-making information, and generating different optimization plans, using the return evaluation model to obtain the predicted return ratio under each optimization plan, and taking the largest predicted return ratio that is greater than the expected return ratio standard as the optimized return ratio, and feeding back the optimization plan corresponding to the optimized return ratio to relevant personnel.

Images