CN114881394A - Big data processing system and method for intelligent evaluation of enterprise growth - Google Patents
Big data processing system and method for intelligent evaluation of enterprise growth Download PDFInfo
- Publication number
- CN114881394A CN114881394A CN202210290091.1A CN202210290091A CN114881394A CN 114881394 A CN114881394 A CN 114881394A CN 202210290091 A CN202210290091 A CN 202210290091A CN 114881394 A CN114881394 A CN 114881394A
- Authority
- CN
- China
- Prior art keywords
- enterprise
- data
- label
- labels
- module
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/063—Operations research, analysis or management
- G06Q10/0639—Performance analysis of employees; Performance analysis of enterprise or organisation operations
- G06Q10/06393—Score-carding, benchmarking or key performance indicator [KPI] analysis
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
- G06F16/21—Design, administration or maintenance of databases
- G06F16/215—Improving data quality; Data cleansing, e.g. de-duplication, removing invalid entries or correcting typographical errors
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
- G06F16/28—Databases characterised by their database models, e.g. relational or object models
- G06F16/284—Relational databases
- G06F16/285—Clustering or classification
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N20/00—Machine learning
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Business, Economics & Management (AREA)
- Databases & Information Systems (AREA)
- Human Resources & Organizations (AREA)
- General Physics & Mathematics (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- General Engineering & Computer Science (AREA)
- Strategic Management (AREA)
- Software Systems (AREA)
- Quality & Reliability (AREA)
- Entrepreneurship & Innovation (AREA)
- Economics (AREA)
- Educational Administration (AREA)
- Development Economics (AREA)
- General Business, Economics & Management (AREA)
- Artificial Intelligence (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Tourism & Hospitality (AREA)
- Evolutionary Computation (AREA)
- Medical Informatics (AREA)
- Operations Research (AREA)
- Computing Systems (AREA)
- Marketing (AREA)
- Mathematical Physics (AREA)
- Game Theory and Decision Science (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
A big data processing system and method for intelligently evaluating enterprise growth comprises an enterprise data preprocessing module, an enterprise value evaluation index management module and an enterprise value evaluation module. The method comprises the following steps of establishing and applying a machine deep learning model of a preprocessing module, and solving the problem that subjective judgment of labels and data is inaccurate; and the enterprise value evaluation index management module freely sets the weighted values of the tags, the algorithms and the labels participating in scoring according to the integrity of the data corresponding to the tags, the enterprise value evaluation module comprises an absolute score evaluation module and a relative score evaluation module, the absolute scores are set and can be sequentially scored according to time, the change conditions of the enterprise over the years are effectively seen, the relative scores are set and can be seen in a certain specific range, and the enterprise obtains the gap relative to the enterprise in other selected ranges based on a gray comparison method.
Description
Technical Field
The invention belongs to the technical field of big data mining and information processing, and particularly relates to a big data processing system and method for intelligently evaluating enterprise growth.
Background
With the development of new economy, innovation and entrepreneurial enterprises are emerging continuously, government and science and technology financial institutions support the innovation and entrepreneurial enterprises and invest more and more, the value evaluation demand of the innovation and entrepreneurial enterprises is greatly increased, but the traditional evaluation means is difficult to adapt to the value evaluation of the current science and technology enterprises due to the factors of insufficient data, asymmetric information, inaccurate subjective judgment, difficult enterprise technical capability evaluation and the like.
Disclosure of Invention
Aiming at the problems, the invention provides a big data processing system and a big data processing method for enterprise growth intelligent evaluation, different evaluation standards are set according to different characteristics of each region and enterprise, each index is scored by utilizing machine deep learning according to enterprise data to obtain an enterprise value evaluation result, the inaccuracy of manual analysis is reduced, an enterprise value ranking list is obtained based on big data analysis, and a decision maker is helped to quickly screen, compare in multiple dimensions and make a precise decision.
The invention is realized by the following technical scheme:
a big data processing system for intelligently evaluating enterprise growth comprises an enterprise data preprocessing module, an enterprise value evaluation index management module and an enterprise value evaluation module; the enterprise data preprocessing module establishes a crotch-type multi-level label pool, classifies various enterprise data according to a machine deep learning model and a keyword and/or high-frequency word cross matching mode, and/or a data coverage rate statistics mode and/or a data coverage range statistics mode, and establishes a multi-dimensional association relation with the crotch-type multi-level label pool; the multidimensional association relation is established to form a multidimensional association enterprise database and an enterprise multidimensional association database, multidimensional association enterprise data of all enterprises in the multidimensional association enterprise database are mainly formed by all labels in the crotch type multistage label pool, all enterprises which accord with label types can be screened out through index labels, multidimensional association enterprise data of the enterprises in the enterprise multidimensional association database are mainly formed by the enterprises, and all labels in the crotch type multistage label pool which corresponds to the enterprises and label data corresponding to all the labels can be screened out through indexing enterprise names and/or enterprise credit codes and/or enterprise random codes which correspond to the enterprises;
the evaluation index management module comprises a tag data integrity display module, an evaluation tag selection submodules, an algorithm selection submodules and a tag index weight selection submodules, wherein the tag data integrity display module calculates the integrity of data under each tag according to the enterprise multidimensional associated data and displays the integrity so as to be used by the evaluation tag selection submodules for tag selection; the evaluation label selection sub-module selects whether the corresponding label participates in score calculation or not according to the integrity of the data under each label displayed by the label data integrity display module; the algorithm selection sub-module selects the evaluation of the relative scores of the corresponding labels to adopt a linear algorithm or an exponential algorithm; the label index weight selection submodules give weights to corresponding indexes of the labels;
the enterprise value evaluation module comprises an absolute score evaluation module and a relative score evaluation module.
Preferably, the absolute score evaluation module compares the multidimensional associated enterprise data of each enterprise corresponding to the labels freely selected by the evaluation label selection sub-module with a preset highest score standard, processes the multidimensional associated enterprise data according to a linear algorithm or an exponential algorithm selected by the algorithm selection sub-module, and multiplies the weight given by the label index weight selection sub-module by the weight given by the label index weight selection sub-module to obtain an absolute score of the enterprise;
preferably, the relative score evaluation module compares the multidimensional associated enterprise data of each enterprise corresponding to the tags freely selected by the evaluation tag selection sub-module with the multidimensional associated enterprise data of all selected enterprises participating in evaluation, processes the multidimensional associated enterprise data according to a linear algorithm or an exponential algorithm selected by the algorithm selection sub-module, and multiplies the weight given by the tag index weight selection sub-module to obtain the relative score of each enterprise in the range of the selected enterprises participating in evaluation.
Preferably, the enterprise multidimensional association data comprises multidimensional screening of the multidimensional association enterprise data in the machine deep learning classification model, and multidimensional association results among enterprises formed after label classification is carried out on each dimension according to the type of the enterprise; each label has corresponding enterprise data, the label comprises a primary label, each primary label has a subordinate secondary label, and the secondary label is a calculation basis of the absolute score and the relative score; the evaluation label selection submodules freely select whether each secondary label participates in score calculation.
Preferably, in the relative-partial evaluation module, the relative basis score of the enterprise is calculated based on the relative basis score of each selected secondary label, wherein the relative basis score of the secondary label of the forward index is calculated in a manner of (20+ (Aij-MinAi)/(MaxAi-MinAi) × 80), and the relative basis score of the secondary label of the reverse index is calculated in a manner of (20+ (Aij-MinAi)/(MaxAi-MinAi) × 80), wherein Aij represents an actual value corresponding to the ith secondary label of the jth enterprise, MaxAi represents a maximum value of the ith index of all enterprises selected to participate in the evaluation, and min represents a minimum value of the ith index of all enterprises selected to participate in the evaluation; the relative scores of the secondary labels are divided into linear relative scores of the secondary labels and exponential relative scores of the secondary labels according to different selected linear algorithms or exponential algorithms, and the linear relative scores of the secondary labels are the relative basic scores of the secondary labels multiplied by the weight of the secondary labels/100; the index relative score of the secondary label is the relative basic score of the secondary label processed according to an index algorithm multiplied by the weight/100 of the secondary label, the relative score of the primary label is the sum of the relative scores of all the secondary labels under the primary label, and the relative score of the enterprise is the sum of the relative scores of all the primary labels.
Preferably, in the absolute score evaluation module, the absolute score of the enterprise is calculated based on the absolute base score of each selected secondary label, wherein the absolute base score of the secondary label of the forward indicator is calculated in a manner of (20+ actual value/highest score standard x 80 corresponding to the secondary label); the absolute basic score of the secondary label of the reverse index is calculated in a mode of (20+ (highest scoring standard-actual value corresponding to the secondary label)/highest scoring standard x 80); the absolute scores of the secondary labels are divided into linear absolute scores of the secondary labels and exponential absolute scores of the secondary labels according to different selected linear algorithms or exponential algorithms, and the linear absolute scores of the secondary labels are the absolute basic scores of the secondary labels multiplied by the weight of the secondary labels per 100; the exponential absolute score of the secondary label is that the absolute basic score of the secondary label is processed according to an exponential algorithm and then multiplied by the weight of the secondary label/100, the absolute score of the primary label is the sum of the absolute scores of all the secondary labels under the primary label, and the absolute score of the enterprise is the sum of the absolute scores of all the primary labels.
Preferably, the primary label comprises one or more of enterprise size and contribution degree, research and development innovation capacity, talent reserve, investment and financing capacity, enterprise qualification, competitive capacity and enterprise credit investigation.
Preferably, the secondary labels of the enterprise scale and the contribution degree comprise one or more of business income, revenue sharing and revenue sharing acceleration, tax payment, number of insured persons, number of subsidiaries and effective patent amount of invention; the secondary labels of talent reserves comprise one or more of the number of major staff, the number of research and development staff, the number of highly skilled talents, the proportion of major staff to total staff, the proportion of research and development staff to total staff, the proportion of highly skilled talent to total staff, the type of talent plan selected by enterprise founders and the number of ten thousand planned talents; the secondary label of the financing investment capacity comprises one or more of participation in industry development or innovation startup fund, bond financing or equity financing or other financing; the secondary label of the enterprise qualification comprises one or more of high-enterprise and continuous confirmation situations, gazelle and continuous confirmation situations and situations for obtaining various levels of rewards; the competitive secondary label comprises one or more of the development trend of the industry, the degree of conformity with the local industry and the number of patent technology competitors of the same type.
Preferably, the forward label is a label representing upward or forward development and growth, and the larger the actual value corresponding to the label is, the better the evaluation is, such as business income; the reverse index refers to an index with a smaller numerical value and a better numerical value, such as a failed patent ratio.
Preferably, the multidimensional associated enterprise data includes data formed by classifying and then summarizing (multidimensional associating) all enterprise original data of an enterprise to enterprise names and/or enterprise credit codes and/or enterprise random codes corresponding to the enterprise, and includes enterprise basic information data, related news report data, technical standard data, intellectual property data, product data and/or industrial field data; the enterprise basic information at least comprises an enterprise name, a service range and a registration address.
Preferably, the enterprise data preprocessing module comprises an enterprise data acquisition submodule, an enterprise data cleaning submodule, a data manual screening submodule, a machine deep learning model establishing submodule and an enterprise label multi-dimensional screening submodule;
the enterprise data acquisition sub-module acquires various enterprise data from various open channels and sends the acquired enterprise data to the enterprise data cleaning sub-module;
the enterprise data cleaning submodule carries out data cleaning operation on enterprise data according to a cleaning principle to obtain multi-dimensional associated enterprise data, and sends the obtained multi-dimensional associated enterprise to the data manual screening submodule and the machine deep learning model building submodule; the cleaning principle comprises a data generation time node, data validity judgment and data integrity judgment;
the data manual screening submodule is used for manually screening the multidimensional associated enterprise data through keyword and/or high-frequency word cross matching, data coverage rate statistics and data coverage range statistics of the data to generate a manual logic target result and sending the manual logic target result to the machine deep learning model building submodule; the artificial logic target result comprises an artificial label classification result and other label classification results of small-range multi-angle data
The machine deep learning model building sub-module takes an artificial logic target result as a comparison group, and labels the multi-dimensional associated enterprise data for a plurality of times through a machine deep learning algorithm until the relative error between the obtained result and the comparison group is smaller than a preset error threshold value, so that a machine deep learning classification basis model conforming to artificial logic is obtained;
the enterprise label multidimensional screening submodule classifies multidimensional associated enterprise data according to the type of the enterprise according to the machine deep learning classification model, and forms the label classification result into enterprise multidimensional associated data; and the multidimensional associated enterprise data obtained after the multidimensional associated enterprise data is processed by the machine deep learning classification model is consistent with the artificial logic target result.
A big data processing method for intelligent evaluation of enterprise growth comprises the following steps,
s1, acquiring various enterprise original data from various open channels;
s2, performing data cleaning on the various enterprise original data to obtain multidimensional associated enterprise data;
s3, manually screening the multidimensional associated enterprise data through keyword and/or high-frequency word cross matching, and/or data coverage rate statistics, and/or data coverage range statistics, and generating a manual logic target result according to the data type;
s4, taking the artificial logic target results as a plurality of comparison groups, and performing label classification machine correction on the multidimensional associated enterprise data for a plurality of times through a machine deep learning algorithm until the relative error between the obtained machine results and the artificial logic target results of the comparison groups is smaller than a preset error threshold value, so as to obtain a machine deep learning classification model conforming to artificial logic;
s5, performing label classification on the multidimensional associated enterprise data according to the type of the enterprise, and forming enterprise multidimensional associated data according to the label classification result; the multidimensional associated enterprise data obtained after the multidimensional associated enterprise data is processed by the machine deep learning classification model is consistent with an artificial logic target result;
s6, calculating the integrity of the data under each label according to the enterprise multidimensional correlation data, selecting the labels participating in score evaluation according to the integrity of the data under each label, and selecting scores for each selected label to perform linear algorithm processing or exponential algorithm processing; respectively calculating absolute scores and relative scores of enterprises, and visually displaying the scores obtained by the enterprises through a radar map from the dimensionality of the labels;
the method comprises the steps of calculating absolute scores and relative scores of enterprises, wherein the absolute scores and the relative scores of the enterprises are related to the number of label dimensions, when the label dimensions are two-dimensional, the labels comprise primary labels and secondary labels, each primary label has a secondary label subordinate to the primary label, and the secondary labels are the basis for calculating the absolute scores and the relative scores.
Compared with the prior art, the invention has the advantages that:
according to the big data processing system and method for enterprise growth intelligent evaluation, an enterprise data acquisition module can perform data entry in a mode of internet data acquisition and enterprise self active declaration, and the problem of information asymmetry is solved; the problem of inaccurate subjective judgment of labels and data is solved by establishing and applying a machine deep learning model; when the related data corresponding to the specific label exists, the score can be scored according to the algorithm, and when the related data corresponding to the label is absent, the machine deep learning model can perform comparative learning according to other information collected from the interconnection night and other enterprise information with data, so that a relatively fair score can be scored for enterprises without specific index values. The absolute scores can be sequentially scored according to the time sequence, so that the change conditions of the enterprises over the years can be effectively seen, the relative scores can be seen in a certain specific range, and the gap of the enterprises in other selected ranges is obtained based on a gray comparison method. And the evaluation index management module freely sets the weighted values of the tags, the algorithms and the labels participating in the scoring according to the integrity of the data corresponding to the tags, meets various requirements under different conditions, does not depend on a mode of holding an expert review meeting, and solves the problems of opaque scoring indexes, poor index adaptability, poor index personalization, insufficient data, asymmetric information, inaccurate subjective judgment, difficult enterprise technical capability evaluation and the like in the prior art. And the intelligent assessment result of the enterprise growth can be more freely and fully displayed through a radar map and the like.
Drawings
FIG. 1 is a flow diagram of a big data processing system and method of the present invention for intelligent assessment of enterprise growth.
Detailed Description
In order to facilitate an understanding of the invention, the invention is described in more detail below with reference to the accompanying drawings and specific examples.
A big data processing system and method for enterprise growth intelligent assessment are disclosed, as shown in FIG. 1, comprising an enterprise data preprocessing module, an enterprise value evaluation index management module and an enterprise value assessment module;
the enterprise data preprocessing module establishes a crotch-type multi-level label pool, classifies various enterprise data according to a machine deep learning model and a keyword and/or high-frequency word cross matching mode, and/or a data coverage rate statistics mode and/or a data coverage range statistics mode, and establishes a multi-dimensional association relation with the crotch-type multi-level label pool; and after the multidimensional association relation is established, a multidimensional association enterprise database which comprises multidimensional association enterprise data of all enterprises which are formed by taking all the labels in the crotch type multistage label pool as main parts and are in accordance with the label types through label data screening, and an enterprise multidimensional association enterprise database which is formed by taking the enterprises as main parts and is used for indexing the enterprise multidimensional association data of the label data of all the labels in the crotch type multistage label pool corresponding to the enterprises through enterprise names and/or enterprise credit codes and/or enterprise random codes corresponding to the enterprises.
The enterprise data preprocessing module comprises an enterprise data acquisition submodule, an enterprise data cleaning submodule, a data manual screening submodule, a machine deep learning model establishing submodule, an enterprise label multi-dimensional screening submodule, an enterprise value evaluation index management module and an enterprise value evaluation module;
the enterprise data acquisition sub-module acquires various enterprise data from various open channels and sends the acquired enterprise data to the enterprise data cleaning module; the enterprise data includes, but is not limited to, enterprise basic information data, related news report data, technical standard data, intellectual property data, product data, and/or industrial domain data. The enterprise basic information at least comprises an enterprise name, a service range and a registration address. The data acquisition sub-module can update the data in real time or capture the data in fixed time or input the data in fixed time so as to update the original data of the enterprise in time.
The enterprise data cleaning submodule carries out data cleaning operation on enterprise data according to a cleaning principle to obtain multi-dimensional associated enterprise data, and sends the obtained multi-dimensional associated enterprise to the data manual screening submodule and the machine deep learning model building submodule; the cleaning principle comprises data generation time nodes, and data validity judgment and data integrity judgment are carried out, namely when data cleaning processing is carried out on government network open data, enterprise interaction data, professional organization data (such as sky eye inspection and the like) and other data from the Internet, at least the data generation time nodes of various enterprise original data need to be captured, and data validity judgment and data integrity judgment are carried out.
The multidimensional associated enterprise data obtained after cleaning comprises data formed by classifying and summarizing (multidimensional associating) all enterprise original data of a certain enterprise to enterprise names and/or enterprise credit codes and/or enterprise random codes corresponding to the enterprise, wherein the data types comprise registered funds, establishment time, personnel scale, enterprise properties, whether gazelle is higher or not, whether enterprise is on the market or not, whether independent angle is beast or not, whether three new plates are provided or not, whether double-creation activities are provided or not, whether a platform is created or not, whether an enterprise carrier is created or not, how many research and development are provided for the industry, total industrial value per year, product sales income, net profit, asset liability rate, income gain tax, cost expense rate, owner equity, technical income, technical service export ratio, high-technology product export ratio, total profit, production profit rate, total production contract contribution amount, and income amount, The province, the city, the district, the city type, the high and new district obtain one or more of the national science and technology rewards, what enterprise qualification is, the latest financing round, the latest financing amount, the maximum financing amount, the latest financing time, the annual R & D ratio, the R & D investment, the R & D personnel ratio, the R & D personnel wage ratio, the patent holding amount, the industry field, the high and new technology field, the national economy industry classification, the strategic emerging industry field and the industry position information. Each data type being a tag.
The data manual screening submodule is used for manually screening the multidimensional associated enterprise data through keyword and/or high-frequency word cross matching, data coverage rate statistics and data coverage range statistics of the data to generate a manual logic target result and sending the manual logic target result to the machine deep learning model building submodule;
the machine deep learning model building sub-module takes an artificial logic target result as a comparison group, and labels the multi-dimensional associated enterprise data for a plurality of times through a machine deep learning algorithm until the relative error between the obtained result and the comparison group is smaller than a preset error threshold value, so that a machine deep learning classification basis model conforming to artificial logic is obtained; the artificial logical target result includes all labels generated according to the data type and all data under the labels. Specifically, if the manual data screening module obtains 1000 tags, each tag has assigned to it a corresponding type of data. The machine deep learning model building module performs machine learning on the data of the class corresponding to the first 500 labels through a machine deep learning algorithm, so that the data can just appear the first 500 labels, a preliminary machine deep learning model is built, then the preliminary machine deep learning model tags the data corresponding to the 600 th labels 501-, and obtaining a machine deep learning classification model which accords with artificial logic.
The enterprise label multidimensional screening submodule classifies multidimensional associated enterprise data according to the type of the enterprise according to the machine deep learning classification model, and forms the label classification result into enterprise multidimensional associated data; the multidimensional associated enterprise data obtained after the multidimensional associated enterprise data is processed by the machine deep learning classification model is consistent with an artificial logic target result; and the enterprise multidimensional association data comprises multidimensional screening on the machine deep learning classification model, and multidimensional association results among enterprises formed after label classification is respectively carried out on each dimension according to the type of the enterprise. The label types in the label classification result comprise one or more of an industry field label, a standing time label, a latest financing round label, a latest financing amount label, a latest financing time label, a personnel scale label, a registered fund label, an enterprise property label, a patent holding amount label, a business income label, a technical income label, an R & D investment label and an R & D proportion label.
The evaluation index management module comprises a tag data integrity display module, an evaluation tag selection submodules, an algorithm selection submodules and a tag index weight selection submodules, wherein the tag data integrity display module calculates the integrity of data under each tag according to the enterprise multidimensional associated data and displays the integrity so as to be used by the evaluation tag selection submodules for tag selection; the evaluation label selection sub-module selects whether the corresponding label participates in score calculation or not according to the integrity of the data under each label displayed by the label data integrity display module; the algorithm selection sub-module selects the evaluation of the relative scores of the corresponding labels to adopt a linear algorithm or an exponential algorithm; the label index weight selection submodules give weights to corresponding indexes of the labels;
the enterprise value evaluation module comprises an absolute score evaluation module and a relative score evaluation module;
the absolute score evaluation module compares the multidimensional associated enterprise data of each enterprise corresponding to the labels freely selected by the evaluation label selection sub-module with a preset highest scoring standard, processes the multidimensional associated enterprise data according to a linear algorithm or an exponential algorithm selected by the algorithm selection sub-module, and multiplies the multidimensional associated enterprise data by the weight given by the label index weight selection sub-module to obtain the absolute score of the enterprise;
and the relative score evaluation module compares the multidimensional associated enterprise data of each enterprise corresponding to the labels freely selected by the evaluation label selection sub-module with the multidimensional associated enterprise data of all selected enterprises participating in evaluation, processes the multidimensional associated enterprise data according to a linear algorithm or an exponential algorithm selected by the algorithm selection sub-module, and multiplies the weight given by the label index weight selection sub-module to obtain the relative score of each enterprise in the enterprise range selected to participate in evaluation.
Example 2
Different from the embodiment, the enterprise multidimensional association data comprises multidimensional screening of the multidimensional association enterprise data in the machine deep learning classification model, and multidimensional association results among enterprises formed after label classification is respectively carried out on each dimension according to the type of the enterprise; each label has corresponding enterprise data, the label comprises a primary label, each primary label has a subordinate secondary label, and the secondary label is a calculation basis of the absolute score and the relative score; the evaluation label selection submodules freely select whether each secondary label participates in score calculation. The primary labels comprise enterprise scale and contribution degree, research and development innovation capacity, talent reserve, investment and financing capacity, enterprise qualification, competitive capacity and enterprise credit investigation. The secondary labels of the enterprise scale and the contribution degree comprise business income, revenue sharing and share rate acceleration, tax payment, number of insured persons, number of subsidiaries and effective patent amount of invention; the secondary labels of talent reserves comprise the number of large talents, the number of research and development personnel, the number of high-skill talents, the proportion of large talents to total employees, the proportion of research and development personnel to total employees, the proportion of high-skill talents to total employees, the type of talent plan selected by enterprise founders and the number of ten-thousand planned talents; the secondary label of the investment and financing capacity comprises participation of industrial development or innovation startup fund, bond financing, equity financing and other financing; the secondary label of the enterprise qualification comprises high-enterprise and continuous confirmation conditions, gazelle and continuous confirmation conditions and conditions for obtaining all levels of rewards; the competitive secondary label comprises the development trend of the industry, the degree of conformity with the local industry and the number of competitors of the similar patent technology. The forward label is a label representing upward or forward development and growth, and the larger the actual value corresponding to the label is, the better the evaluation is, such as business income; the reverse index refers to an index with a smaller numerical value and a better numerical value, such as a failed patent ratio.
In the relative evaluation module, the relative basis score of enterprises is calculated based on the relative basis score of each selected secondary label, wherein the relative basis score of the secondary label of the forward index is calculated in a mode of (20+ (Aij-MinAi)/(Maxai-MinAi) x 80), the relative basis score of the secondary label of the reverse index is calculated in a mode of (20+ (Aij-MinAi)/(Maxai-MinAi) x 80), wherein Aij represents an actual value corresponding to the ith secondary label of the jth enterprise, Maxai represents the maximum value of the ith index of all enterprises selected to participate in evaluation, and Minai represents the minimum value of the ith index of all enterprises selected to participate in evaluation; the relative scores of the secondary labels are divided into linear relative scores of the secondary labels and exponential relative scores of the secondary labels according to different selected linear algorithms or exponential algorithms, and the linear relative scores of the secondary labels are the relative basic scores of the secondary labels multiplied by the weight of the secondary labels/100; the index relative score of the secondary label is the relative basic score of the secondary label processed according to an index algorithm multiplied by the weight/100 of the secondary label, the relative score of the primary label is the sum of the relative scores of all the secondary labels under the primary label, and the relative score of the enterprise is the sum of the relative scores of all the primary labels.
In the absolute score evaluation module, the absolute score of the enterprise is calculated based on the absolute basic score of each selected secondary label, wherein the absolute basic score of the secondary label of the forward index is calculated in a mode of (20+ actual value/highest score standard x 80) corresponding to the secondary label; the absolute basic score of the secondary label of the reverse index is calculated in a mode of (20+ (highest scoring standard-actual value corresponding to the secondary label)/highest scoring standard x 80); the absolute scores of the secondary labels are divided into linear absolute scores of the secondary labels and exponential absolute scores of the secondary labels according to different selected linear algorithms or exponential algorithms, and the linear absolute scores of the secondary labels are the absolute basic scores of the secondary labels multiplied by the weight of the secondary labels per 100; the exponential absolute score of the secondary label is that the absolute basic score of the secondary label is processed according to an exponential algorithm and then multiplied by the weight of the secondary label/100, the absolute score of the primary label is the sum of the absolute scores of all the secondary labels under the primary label, and the absolute score of the enterprise is the sum of the absolute scores of all the primary labels.
A big data processing method for intelligent evaluation of enterprise growth comprises the following steps,
s1, acquiring various enterprise original data from various open channels;
s2, performing data cleaning on the various enterprise original data to obtain multidimensional associated enterprise data;
s3, manually screening the multidimensional associated enterprise data through keyword and/or high-frequency word cross matching, and/or data coverage rate statistics, and/or data coverage range statistics, and generating a manual logic target result according to the data type;
s4, taking the artificial logic target results as a plurality of comparison groups, and performing label classification machine correction on the multidimensional associated enterprise data for a plurality of times through a machine deep learning algorithm until the relative error between the obtained machine results and the artificial logic target results of the comparison groups is smaller than a preset error threshold value, so as to obtain a machine deep learning classification model conforming to artificial logic;
s5, performing label classification on the multidimensional associated enterprise data according to the type of the enterprise, and forming enterprise multidimensional associated data according to the label classification result; the multidimensional associated enterprise data obtained after the multidimensional associated enterprise data is processed by the machine deep learning classification model is consistent with an artificial logic target result;
s6, calculating the integrity of the data under each label according to the enterprise multidimensional correlation data, selecting the labels participating in score evaluation according to the integrity of the data under each label, and selecting scores for each selected label to perform linear algorithm processing or exponential algorithm processing; respectively calculating absolute scores and relative scores of enterprises, and visually displaying the scores obtained by the enterprises through a radar map from the dimensionality of the labels;
the method comprises the steps of calculating absolute scores and relative scores of enterprises, wherein the absolute scores and the relative scores of the enterprises are related to the number of label dimensions, when the label dimensions are two-dimensional, the labels comprise primary labels and secondary labels, each primary label has a secondary label subordinate to the primary label, and the secondary labels are the basis for calculating the absolute scores and the relative scores.
It should be noted that the above-described embodiments may enable those skilled in the art to more fully understand the present invention, but do not limit the present invention in any way. Therefore, although the present invention has been described in detail with reference to the drawings and examples, it will be understood by those skilled in the art that various changes and modifications can be made therein without departing from the spirit and scope of the invention.
Claims (10)
1. A big data processing system for intelligently evaluating enterprise growth is characterized by comprising an enterprise data preprocessing module, an enterprise value evaluation index management module and an enterprise value evaluation module;
the enterprise data preprocessing module establishes a crotch-type multi-level label pool, classifies various enterprise data according to a machine deep learning model and a keyword and/or high-frequency word cross matching mode, and/or a data coverage rate statistics mode and/or a data coverage range statistics mode, and establishes a multi-dimensional association relation with the crotch-type multi-level label pool; the multidimensional association relation is established to form a multidimensional association enterprise database and an enterprise multidimensional association database, multidimensional association enterprise data of all enterprises in the multidimensional association enterprise database are mainly formed by all labels in the crotch type multistage label pool, all enterprises which accord with label types can be screened out through index labels, multidimensional association enterprise data of the enterprises in the enterprise multidimensional association database are mainly formed by the enterprises, and all labels in the crotch type multistage label pool which corresponds to the enterprises and label data corresponding to all the labels can be screened out through indexing enterprise names and/or enterprise credit codes and/or enterprise random codes which correspond to the enterprises;
the evaluation index management module comprises a tag data integrity display module, an evaluation tag selection submodules, an algorithm selection submodules and a tag index weight selection submodules, wherein the tag data integrity display module calculates the integrity of data under each tag according to the enterprise multidimensional associated data and displays the integrity so as to be used by the evaluation tag selection submodules for tag selection; the evaluation label selection sub-module selects whether the corresponding label participates in score calculation or not according to the integrity of the data under each label displayed by the label data integrity display module; the algorithm selection sub-module selects the evaluation of the relative scores of the corresponding labels to adopt a linear algorithm or an exponential algorithm; the label index weight selection submodules give weights to corresponding indexes of the labels;
the enterprise value evaluation module comprises an absolute score evaluation module and a relative score evaluation module.
2. The big data processing system for intelligent enterprise growth assessment according to claim 1, wherein the absolute score assessment module compares the multidimensional associated enterprise data of each enterprise corresponding to the labels freely selected by the assessment label selection sub-module with a preset highest scoring standard, processes the multidimensional associated enterprise data according to a linear algorithm or an exponential algorithm selected by the algorithm selection sub-module, and multiplies the linear algorithm or the exponential algorithm by the weight given by the label index weight selection sub-module to obtain the absolute score of the enterprise.
3. The big data processing system for intelligent enterprise growth assessment according to claim 1 or 2, wherein the relative score assessment module compares the multidimensional associated enterprise data of each enterprise corresponding to a plurality of tags freely selected by the assessment tag selection sub-module with the multidimensional associated enterprise data of all enterprises selected to participate in the assessment, and then processes the multidimensional associated enterprise data according to the linear algorithm or the exponential algorithm selected by the algorithm selection sub-module and multiplies the weights given by the tag index weight selection sub-module to obtain the relative scores of each enterprise within the enterprise selected to participate in the assessment.
4. The big data processing system for intelligent enterprise growth assessment according to any one of claims 1-3, wherein the enterprise multidimensional association data comprises multidimensional screening of the multidimensional association enterprise data in the machine deep learning classification model, and multidimensional association results between enterprises formed after label classification is performed on each dimension according to the type of the enterprise; each label has corresponding enterprise data, the label comprises a primary label, each primary label has a subordinate secondary label, and the secondary label is a calculation basis of the absolute score and the relative score; the evaluation label selection submodules freely select whether each secondary label participates in score calculation.
5. The big data processing system for intelligent enterprise growth evaluation according to claim 3 or 4, wherein in the relative assessment module, the relative enterprise score is calculated based on the relative base score of each selected secondary label, wherein the relative base score of the secondary label of the forward index is calculated as (20+ (Aij-Minai)/(Maxai-Minai) x 80), and the relative base score of the secondary label of the reverse index is calculated as (20+ (Aij-Minai)/(Maxai-Minai) x 80), wherein Aij represents the actual value corresponding to the ith enterprise secondary label, Maxai represents the maximum value of the ith enterprise index selected to participate in the evaluation, and Minai represents the minimum value of the ith enterprise index selected to participate in the evaluation; the relative scores of the secondary labels are divided into linear relative scores of the secondary labels and exponential relative scores of the secondary labels according to different selected linear algorithms or exponential algorithms, and the linear relative scores of the secondary labels are the relative basic scores of the secondary labels multiplied by the weight of the secondary labels/100; the index relative score of the secondary label is processed according to an index algorithm and multiplied by the weight/100 of the secondary label, the relative score of the primary label is the sum of the relative scores of all the secondary labels under the primary label, and the relative score of the enterprise is the sum of the relative scores of all the primary labels.
6. The big data processing system for intelligent enterprise growth assessment according to claim 2 or 4, wherein the absolute score of the enterprise in said absolute score assessment module is calculated based on the absolute base score of each selected secondary label, wherein the absolute base score of said secondary label of the forward direction index is calculated in the way of (20+ actual value/highest score standard x 80 for the secondary label); the absolute basic score of the secondary label of the reverse index is calculated in a mode of (20+ (highest scoring standard-actual value corresponding to the secondary label)/highest scoring standard x 80); the absolute scores of the secondary labels are divided into linear absolute scores of the secondary labels and exponential absolute scores of the secondary labels according to different selected linear algorithms or exponential algorithms, and the linear absolute scores of the secondary labels are the absolute basic scores of the secondary labels multiplied by the weight of the secondary labels per 100; the exponential absolute score of the secondary label is that the absolute basic score of the secondary label is processed according to an exponential algorithm and then multiplied by the weight of the secondary label/100, the absolute score of the primary label is the sum of the absolute scores of all the secondary labels under the primary label, and the absolute score of the enterprise is the sum of the absolute scores of all the primary labels.
7. The big data processing system for intelligent assessment of enterprise growth according to claim 4, wherein said primary label comprises one or more of enterprise size and contribution, development and innovation capability, talent reserve, investment and financing capability, enterprise qualification, competitive capability, and enterprise credit investigation; the secondary labels of the enterprise scale and the contribution degree comprise one or more of business income, revenue sharing and share rate acceleration, tax payment, number of insured persons, number of subsidiaries and effective patent amount of invention; the secondary labels of talent reserves comprise one or more of the number of major staff, the number of research and development staff, the number of highly skilled talents, the proportion of major staff to total staff, the proportion of research and development staff to total staff, the proportion of highly skilled talent to total staff, the type of talent plan selected by enterprise founders and the number of ten thousand planned talents; the secondary label of the financing investment capacity comprises one or more of participation in industry development or innovation startup fund, bond financing or equity financing or other financing; the secondary label of the enterprise qualification comprises one or more of high-enterprise and continuous confirmation situations, gazelle and continuous confirmation situations and situations for obtaining various levels of rewards; the competitive secondary label comprises one or more of the development trend of the industry, the degree of conformity with the local industry and the number of patent technology competitors of the same type.
8. A big data processing system for intelligent assessment of enterprise growth according to claim 5 or 6, wherein said forward tags are tags representing upward or forward development and growth, and the tags correspond to actual values with better evaluation, such as business income; the reverse index refers to an index with a smaller numerical value and a better numerical value, such as a failed patent ratio.
9. The big data processing system for intelligent enterprise growth assessment according to claim 1, wherein the multidimensional associated enterprise data includes data formed by classifying all enterprise raw data of an enterprise and then summarizing the classified data to enterprise names and/or enterprise credit codes and/or enterprise random codes corresponding to the enterprise, and includes enterprise basic information data, related news report data, technical standard data, intellectual property data, product data, and/or industrial field data; the enterprise basic information at least comprises an enterprise name, a service range and a registration address;
the enterprise data preprocessing module comprises an enterprise data acquisition submodule, an enterprise data cleaning submodule, a data manual screening submodule, a machine deep learning model establishing submodule and an enterprise label multi-dimensional screening submodule;
the enterprise data acquisition sub-module acquires various enterprise data from various open channels and sends the acquired enterprise data to the enterprise data cleaning sub-module;
the enterprise data cleaning submodule carries out data cleaning operation on enterprise data according to a cleaning principle to obtain multi-dimensional associated enterprise data, and sends the obtained multi-dimensional associated enterprise to the data manual screening submodule and the machine deep learning model building submodule; the cleaning principle comprises a data generation time node, data validity judgment and data integrity judgment;
the data manual screening submodule is used for manually screening the multidimensional associated enterprise data through keyword and/or high-frequency word cross matching, data coverage rate statistics and data coverage range statistics of the data to generate a manual logic target result and sending the manual logic target result to the machine deep learning model building submodule; the artificial logic target result comprises an artificial label classification result and other label classification results of small-range multi-angle data
The machine deep learning model building sub-module takes an artificial logic target result as a comparison group, and labels the multi-dimensional associated enterprise data for a plurality of times through a machine deep learning algorithm until the relative error between the obtained result and the comparison group is smaller than a preset error threshold value, so that a machine deep learning classification basis model conforming to artificial logic is obtained;
the enterprise label multidimensional screening submodule classifies multidimensional associated enterprise data according to the type of the enterprise according to the machine deep learning classification model, and forms the label classification result into enterprise multidimensional associated data; and the multidimensional associated enterprise data obtained after the multidimensional associated enterprise data is processed by the machine deep learning classification model is consistent with the artificial logic target result.
10. A big data processing method for intelligent evaluation of enterprise growth is characterized by comprising the following steps,
s1, acquiring various enterprise original data from various open channels;
s2, performing data cleaning on the various enterprise original data to obtain multidimensional associated enterprise data;
s3, manually screening the multidimensional associated enterprise data through keyword and/or high-frequency word cross matching, and/or data coverage rate statistics, and/or data coverage range statistics, and generating a manual logic target result according to the data type;
s4, taking the artificial logic target results as a plurality of comparison groups, and performing label classification machine correction on the multidimensional associated enterprise data for a plurality of times through a machine deep learning algorithm until the relative error between the obtained machine results and the artificial logic target results of the comparison groups is smaller than a preset error threshold value, so as to obtain a machine deep learning classification model conforming to artificial logic;
s5, performing label classification on the multidimensional associated enterprise data according to the type of the enterprise, and forming enterprise multidimensional associated data according to the label classification result; the multidimensional associated enterprise data obtained after being processed by the machine deep learning classification model is consistent with an artificial logic target result;
s6, calculating the integrity of the data under each label according to the enterprise multidimensional correlation data, selecting the labels participating in score evaluation according to the integrity of the data under each label, and selecting scores for each selected label to perform linear algorithm processing or exponential algorithm processing; respectively calculating absolute scores and relative scores of enterprises, and visually displaying the scores obtained by the enterprises through a radar map from the dimensionality of the labels;
the method comprises the steps of calculating absolute scores and relative scores of enterprises, wherein the absolute scores and the relative scores of the enterprises are related to the number of label dimensions, when the label dimensions are two-dimensional, the labels comprise primary labels and secondary labels, each primary label has a secondary label subordinate to the primary label, and the secondary labels are the basis for calculating the absolute scores and the relative scores.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210290091.1A CN114881394A (en) | 2022-03-23 | 2022-03-23 | Big data processing system and method for intelligent evaluation of enterprise growth |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210290091.1A CN114881394A (en) | 2022-03-23 | 2022-03-23 | Big data processing system and method for intelligent evaluation of enterprise growth |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN114881394A true CN114881394A (en) | 2022-08-09 |
Family
ID=82667417
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210290091.1A Pending CN114881394A (en) | 2022-03-23 | 2022-03-23 | Big data processing system and method for intelligent evaluation of enterprise growth |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114881394A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115713399A (en) * | 2022-09-28 | 2023-02-24 | 睿智合创(北京)科技有限公司 | User credit assessment system combined with third-party data source |
-
2022
- 2022-03-23 CN CN202210290091.1A patent/CN114881394A/en active Pending
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115713399A (en) * | 2022-09-28 | 2023-02-24 | 睿智合创(北京)科技有限公司 | User credit assessment system combined with third-party data source |
| CN115713399B (en) * | 2022-09-28 | 2023-10-20 | 睿智合创(北京)科技有限公司 | User credit evaluation system combined with third-party data source |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Aruldoss et al. | A survey on multi criteria decision making methods and its applications | |
| Kronthaler | Economic capability of East German regions: Results of a cluster analysis | |
| Zubairu et al. | Supply chain strategies as drivers of financial performance in liquefied natural gas networks | |
| Yılmaz et al. | Assessment of sustainability performances of banks by TOPSIS method and balanced scorecard approach | |
| Shahbazi Alenjagh | Performance evaluation and ranking of insurance companies in Tehran Stock Exchange by financial ratios using ANP and PROMETHEE | |
| Ajripour et al. | A model for organization performance management applying MCDM and BSC: a case study | |
| Valmohammadi et al. | Six sigma project selection using the hybrid approach FAHP-FTOPSIS and grey relational analysis model | |
| CN115063181A (en) | Entrepreneurship analysis matching financing issuing method | |
| Colapinto et al. | Goal programming for financial portfolio management: a state-of-the-art review | |
| Pan | The use of genetic programming for the construction of a financial management model in an enterprise | |
| Abeysekara | Application of fuzzy set theory to evaluate large scale transport infrastructure risk assessment and application of best practices for risk management | |
| CN114881394A (en) | Big data processing system and method for intelligent evaluation of enterprise growth | |
| Wang et al. | Applying TOPSIS method to evaluate the business operation performance of Vietnam listing securities companies | |
| Amiran et al. | A fuzzy MCDM approach for evaluating steel industry performance based on balanced scorecard: A case in Iran | |
| Bazrkar et al. | IDENTIFYING AND SELECTING THE STRATEGIC PROCESS USING THE CROSS-EFFICIENCY APPROACH BASED ON SATISFACTION LEVEL AND EXTENDDED BALANCED SCORECARD. | |
| Petersone et al. | A Data-Driven Framework for Identifying Investment Opportunities in Private Equity | |
| Akbari et al. | A mixed system of network data envelopment analysis to evaluate the performance of bank branches: an illustration with Iranian banks | |
| Boďa et al. | Spatial aspect in bank branch performance management | |
| Hosein et al. | Evaluating and ranking performance by combination model of balanced scorecard and ariadne uncertain estimate | |
| Witkowska | Discrete choice model application to the credit risk evaluation | |
| Meredith | Accounting for the digital economy: time for a fresh approach | |
| Mohiuddin et al. | Firm Level Performance of Offshore Outsourcing Strategy of Manufacturing Entreprises: A Research Agenda | |
| Bakhshi et al. | Developing a hybrid approach to credit priority based on accounting variables (using analytical network process (ANP) and multi-criteria decision-making) | |
| Lawrence et al. | Balance Scorecard Strategy and Financial Performance of Selected Food and Beverage Industry in Nigeria | |
| Li et al. | How to predict the entry mode of international construction contractor: a case-based reasoning approach |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |