An evaluation method and system for enterprise innovation and growth capabilities based on big data analysis

Abstract

The evaluation method and system of enterprise innovation and growth ability based on big data analysis provided by the present invention relate to the field of quantitative evaluation, and the method includes: collecting sample data of each enterprise dimension of sample enterprises, cleaning, calculating and labeling the sample data; training the indicator weight model with the cleaned and classified sample data to obtain the relative importance score of the indicator item and the priority diagram to calculate the final weight value of each indicator item; calculating the final evaluation score of each enterprise to be evaluated according to the enterprise data of each enterprise to be evaluated combined with the final weight value of each indicator item; wherein the enterprise dimensions include industry and commerce, employees, qualifications, intellectual property rights, projects, financing loans, assets and R&D expenses. The present invention accurately measures the growth ability of enterprises by accurately and reliably quantifying and evaluating the intangible assets of technology-based enterprises, including innovation capabilities and technology projects, and calculating the proportion weight, providing data support for the future development planning and decision-making of enterprises.

Description

Technical Field

[0001] The present invention relates to the field of quantitative evaluation, and in particular to a method and system for evaluating the innovation and growth capabilities of an enterprise based on big data analysis. Background Art

[0002] The traditional evaluation method for corporate innovation and growth capabilities based on big data analysis is mainly financial analysis, which analyzes the company's financial data, business model and business plan, such as assets and liabilities, sales profit, etc. This type of corporate growth capability evaluation method can only be evaluated or updated when the company is financing, auditing or publishing financial reports. It is generally aimed at larger companies or companies that have been established for a long time. For entrepreneurial technology companies that have been established for a short time, their characteristics are short operating time, less financial data, and a large proportion of corporate intangible assets. It is not applicable to such companies to use financial analysis to evaluate their growth capabilities.

[0003] Another corporate evaluation technology is corporate portrait, but corporate portrait only displays the company's existing data from different dimensions, without conducting in-depth analysis or quantitative evaluation of the company's business operations.

[0004] Since it is impossible to accurately and reliably assess the growth capabilities of technology-based SMEs with a relatively large proportion of intangible assets, it will be very difficult for such enterprises to obtain financing loans, policy support, etc., which is not conducive to the development and growth of the enterprises and technological innovation. Summary of the invention

[0005] The purpose of the present invention is to provide an evaluation method and system for enterprise innovation and growth capabilities based on big data analysis. By quantitatively evaluating the unquantifiable innovation capabilities and results transformation capabilities in enterprise operations, enterprises can make relevant enhancements based on the evaluation results of each dimension. When implementing policies or corporate investment and financing, more appropriate data support and investment projects can be obtained based on this evaluation structure.

[0006] To achieve the above objectives, the present invention proposes the following technical solution: a method for evaluating enterprise innovation and growth capabilities based on big data analysis, comprising:

[0007] Collect sample data of each enterprise dimension of sample enterprises, and clean, calculate and label the sample data according to time series; the enterprise dimensions include industry and commerce, employees, qualifications, intellectual property rights, projects, financing loans, assets and R&D expenses;

[0008] The cleaned sample data and label classification of sample enterprises are used to train the indicator weight model, so as to obtain the relative importance scores and priority diagrams of the indicator items according to the indicator weight model, and then calculate the final weight value of each indicator item;

[0009] The final evaluation score and ranking of each indicator item of each enterprise to be evaluated are calculated based on the enterprise data of each enterprise to be evaluated and the final weight value of each indicator item; among which, the enterprise data is the sample data of each enterprise dimension after data cleaning of the enterprise to be evaluated.

[0010] Furthermore, the process of cleaning, calculating and labeling the sample data according to the time series includes:

[0011] Obtain the business information of the sample enterprises, as well as the number of researchers and their educational background, qualification certification information, the number of Class I intellectual property rights, the number of Class II intellectual property rights, project application information, loan amount, financing data, R&D expenses and technical contract information in each of the past three years. The loan amount includes the maximum loan amount, the minimum loan amount, the average loan amount and the total loan amount.

[0012] Clean and normalize the above-obtained data, replace the Chinese content with digital representation, and obtain numerical enterprise data;

[0013] Label and classify the numerical enterprise data to determine the indicators, put the data of each indicator into the object of the enterprise data matrix, and calculate the data value of the sample enterprise under each indicator;

[0014] Calculate the sum, average, maximum, minimum, 25% and 75% values ​​of the data values ​​of all sample enterprises under each indicator, and handle abnormal data values.

[0015] Furthermore, the calculation process of the final weight value of the indicator item is:

[0016] Predefine a multi-level data classification label as a calculation indicator item;

[0017] Receive all sample data of sample enterprises, use the gradient boosting decision tree algorithm of the machine learning xgboost library to extract the features of each sample data and classify, sort and group them, combine the input calculation index items to perform softmax loss function plus regularization term calculation, and obtain the relative importance score and priority diagram of each index item one by one;

[0018] The average of the relative importance score of the indicator item and the preset corresponding indicator item proportion score is taken as the final weight value of the indicator item.

[0019] Furthermore, the calculation process of the final evaluation score of each indicator item of each enterprise to be evaluated is as follows:

[0020] Obtain sample data of each enterprise dimension of each enterprise to be evaluated and perform data processing to obtain an enterprise data matrix of each enterprise to be evaluated;

[0021] The index items calculated by the index weight model are divided into five groups of evaluation dimensions, including endogenous innovation, collaborative investment, management and operation capabilities, achievement transformation capabilities and sustainable growth capabilities. The corresponding value of any index item in each group of evaluation dimensions in the enterprise data matrix is ​​multiplied by its corresponding final weight value to obtain the evaluation score of the corresponding index item;

[0022] According to the normal distribution characteristics, the scores of each group of evaluation dimensions are converted into a 100-point system as the final evaluation score of the enterprise indicator item.

[0023] Furthermore, the index items of the sample enterprises include the total number of employees, the proportion of researchers, the proportion of college degree and above, the proportion of postgraduates, the number of enterprise qualifications, industry-university-research qualifications, the number of subordinate institutions, the number of R&D activities, the proportion of R&D expenses, the enterprise's R&D expenses in the previous year, the R&D expenses last year, the project establishment expenses in the previous year, the project establishment expenses last year, the maximum project amount, the minimum project amount, the average project amount, the total project amount, the total number of approved projects, the total number of applied projects, the total amount of equity investment, the total number of equity investments, the total amount of debt in the previous year, the total amount of debt last year, the maximum amount of debt, the minimum amount of debt, the average amount of debt, the total amount of debt, the total number of debts obtained, the total assets in the previous year, the total assets last year, the net assets in the previous year, the net assets last year, the main business income in the previous year, the main business income last year, the cost and expenses in the previous year, the cost and expenses last year, the net profit in the previous year, the net profit last year, Profits and taxes paid in the previous year, profits and taxes paid last year, proportion of main business in the previous year, proportion of main business last year, net profit margin in the previous year, net profit margin last year, management certification, Class I intellectual property rights, Class II intellectual property rights, number of technology contracts in the previous year, number of technology contracts last year, maximum technology transaction amount of technology contracts, minimum technology transaction amount of technology contracts, average technology transaction amount of technology contracts, total technology transaction amount of technology contracts, total number of technology transaction amounts of technology contracts, number of national science and technology awards, number of standards formed, number of scientific and technological achievements transformed, number of high-tech products, high-tech income, total asset growth rate in the previous year, net asset growth rate, main business income growth rate, cost and expense change rate, net profit growth rate, profit and tax paid growth rate, R&D expense growth rate, main business proportion change rate, net profit margin growth rate, R&D expense proportion growth rate, intellectual property number growth rate and R&D personnel proportion growth rate.

[0024] Furthermore, the method for processing abnormal data values ​​is: for data values ​​of proportion type data that are not in the range of 0 to 1, when it is greater than 1, it is set to 75% of the corresponding item, and when it is less than 1, it is set to 25% of the corresponding item; for data with a growth rate greater than 20, it is set to 75% of the corresponding item, and for data with a growth rate less than 20, it is set to 25% of the corresponding item.

[0025] Another solution of the present invention is to disclose an evaluation system for enterprise innovation and growth capabilities based on big data analysis, the system comprising:

[0026] The enterprise data collection and processing module is used to collect sample data of various enterprise dimensions of sample enterprises, and clean, calculate and label the sample data according to the time series; the enterprise dimensions include industry and commerce, employees, qualifications, intellectual property rights, projects, financing loans, assets and R&D expenses;

[0027] The indicator weight module is used to train the indicator weight model based on the sample data and label classification after cleaning of the sample enterprises, so as to obtain the relative importance scores and priority diagrams of the indicator items according to the indicator weight model, and then calculate the final weight value of each indicator item;

[0028] The score assignment module is used to calculate the final evaluation score and ranking of each indicator item of each enterprise to be evaluated based on the enterprise data of each enterprise to be evaluated and the final weight value of each indicator item.

[0029] Furthermore, the enterprise data collection and processing module includes:

[0030] The first acquisition unit is used to obtain the industrial and commercial information of the sample enterprises, as well as the number of researchers and the educational background ratio, qualification certification information, the number of Class I intellectual property rights, the number of Class II intellectual property rights, project application information, loan amount, financing data, R&D expenses and technical contract information in each of the past three years. The loan amount includes the maximum loan amount, the minimum loan amount, the average loan amount and the total loan amount.

[0031] A cleaning and normalization unit is used to clean and normalize the data acquired by the acquisition unit, including deleting duplicate data, evaluating irrelevant data, and replacing Chinese content with digital representation to obtain numerical enterprise data;

[0032] The classification unit is used to classify the numerical enterprise data into labels to determine each indicator item, put the data of each indicator item into the object of the enterprise data matrix, and calculate the data value of the sample enterprise under each indicator item respectively;

[0033] The calculation unit is used to calculate the sum, average, maximum, minimum, 25% and 75% values ​​of the data values ​​of all sample enterprises under each indicator item, and process abnormal data values.

[0034] Furthermore, the indicator weight module includes:

[0035] A predefined unit is used to predefine a multi-level data classification label as a calculation indicator item;

[0036] The machine learning unit is used to receive all sample data of sample enterprises, use the gradient boosting decision tree algorithm of the machine learning xgboost library to extract the features of each sample data and classify, sort and group them, and perform softmax loss function plus regularization calculation on the input calculation index items to obtain the relative importance score and priority diagram of each index item one by one;

[0037] The weight calculation unit is used to calculate the average of the relative importance score of the indicator item and the preset corresponding indicator item proportion score, and use the average as the final weight value of the indicator item.

[0038] Furthermore, the score assignment module includes:

[0039] The second acquisition unit is used to acquire sample data of each enterprise dimension of each enterprise to be evaluated and perform data processing to obtain an enterprise data matrix of each enterprise to be evaluated;

[0040] Divide the evaluation units to divide the indicators used to calculate the indicator weight model into five groups of evaluation dimensions, including endogenous innovation, collaborative investment, management and operation capabilities, achievement transformation capabilities, and sustainable growth capabilities. Perform matrix dot multiplication on the corresponding value of any indicator item in each group of evaluation dimensions in the enterprise data matrix and its corresponding final weight value to obtain the evaluation score of the corresponding indicator item;

[0041] The score conversion unit is used to convert the scores of each group of evaluation dimensions into a 100-point system according to the normal distribution characteristics as the final evaluation score of the enterprise indicator item.

[0042] It can be seen from the above technical solutions that the technical solutions of the present invention have the following beneficial effects:

[0043] The evaluation method and system of enterprise innovation and growth capability based on big data analysis disclosed by the present invention not only considers the financial indicators of enterprises, but also accurately and reliably measures the growth capability of enterprises by accurately and quantitatively evaluating the intangible assets of enterprises, such as innovation capability and scientific and technological projects. Specifically, the method includes: collecting sample data of each enterprise dimension of sample enterprises, cleaning, calculating and labeling the sample data according to time series, wherein the enterprise dimensions include industry and commerce, employees, qualifications, intellectual property rights, projects, financing loans, assets and R&D expenses; using the sample data and label classification after cleaning of sample enterprises, training the indicator weight model, so as to obtain the relative importance score and priority diagram of the indicator items according to the indicator weight model, and then calculating the final weight value of each indicator item; calculating the final evaluation score of each indicator item of each enterprise to be evaluated according to the enterprise data of each enterprise to be evaluated combined with the final weight value of each indicator item; the final evaluation score revolves around five aspects, namely, endogenous innovation, collaborative investment attraction, management and operation capability, achievement transformation capability and sustainable growth capability.

[0044] The final evaluation score of the enterprise obtained by the present invention is more suitable for evaluating the growth capacity of small and medium-sized enterprises than the financial analysis method that focuses on financial indicators, and provides data support for the company's future development planning, policy adaptation and decision-making; and compared with the corporate portrait, it can not only display the company's development data from different dimensions, but also can more deeply analyze the impact of the company's operating behavior and innovation ability on the company's growth ability.

[0045] It should be appreciated that all combinations of the foregoing concepts, as well as additional concepts described in greater detail below, may be considered to be part of the inventive subject matter of the present disclosure, provided such concepts are not mutually inconsistent.

[0046] The foregoing and other aspects, embodiments and features of the present invention can be more fully understood from the following description in conjunction with the accompanying drawings. Other additional aspects of the present invention, such as the features and / or beneficial effects of the exemplary embodiments, will be apparent from the following description or learned from the practice of the specific embodiments according to the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] The drawings are not intended to be drawn to scale. In the drawings, each identical or nearly identical component shown in various figures may be represented by the same reference numeral. For clarity, not every component is labeled in every figure. Embodiments of various aspects of the present invention will now be described by way of example and with reference to the accompanying drawings, in which:

[0048] Figure 1 A flow chart of the method for evaluating the innovation and growth capabilities of an enterprise based on big data analysis according to the present invention;

[0049] Figure 2 Prepare a flow chart for sample data cleaning in the embodiment;

[0050] Figure 3 The sample data cleaning process in the embodiment Figure 1 ;

[0051] Figure 4 The sample data cleaning process in the embodiment Figure 2 ;

[0052] Figure 5 The weight of each evaluation dimension for generating the evaluation index in the embodiment. DETAILED DESCRIPTION

[0053] In order to make the purpose, technical solution and advantages of the embodiment of the present invention clearer, the technical solution of the embodiment of the present invention will be clearly and completely described below in conjunction with the drawings of the embodiment of the present invention. Obviously, the described embodiment is a part of the embodiment of the present invention, not all of the embodiments. Based on the described embodiment of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention. Unless otherwise defined, the technical terms or scientific terms used herein should be the common meaning understood by people with general skills in the field to which the present invention belongs.

[0054] The words "first", "second" and similar words used in the patent application specification and claims of the present invention do not indicate any order, quantity or importance, but are only used to distinguish different components. Similarly, unless the context clearly indicates otherwise, the singular form of "a", "an" or "the" and other similar words do not indicate a quantitative limitation, but indicate the existence of at least one. "Include" or "comprises" and other similar words mean that the elements or objects appearing before "include" or "comprises" include the features, wholes, steps, operations, elements and / or components listed after "include" or "comprises", and do not exclude the existence or addition of one or more other features, wholes, steps, operations, elements, components and / or their collections.

[0055] Based on the existing technology, the evaluation method for enterprise growth ability is mainly financial analysis method, which focuses on the financial indicators of enterprises. As a result, the enterprise growth ability evaluated by this method can only be used when the enterprise is financing, auditing or publishing financial reports, and it can only be applied to large-scale or long-established enterprises. It is not applicable to small and medium-sized enterprises and entrepreneurial technology enterprises with a short establishment time. The reason is that the innovation ability and achievement transformation ability of such enterprises in their business activities are not quantitatively evaluated in this evaluation method. Therefore, the present invention aims to propose an evaluation method and system for enterprise innovation and growth ability based on big data analysis. In addition to common indicators such as financial indicators, the non-quantifiable innovation ability and achievement transformation ability of the enterprise's intangible assets are also quantitatively evaluated, accurately and reliably evaluating the enterprise's growth ability, and providing data support for the future development and decision-making of the enterprise.

[0056] The following is a further detailed introduction to the method and system for evaluating enterprise innovation and growth capabilities based on big data analysis disclosed in the present invention in conjunction with the embodiments shown in the accompanying drawings.

[0057] Combination Figure 1 As shown in the flowchart, the evaluation method of enterprise innovation and growth capability based on big data analysis disclosed in the embodiment of the present application includes the following steps when executed:

[0058] Step S102, collect sample data of each enterprise dimension of sample enterprises, and perform data cleaning, calculation and label classification on the sample data according to time series; wherein the enterprise dimensions include industry and commerce, employees, qualifications, intellectual property rights, projects, financing loans, assets and R&D expenses;

[0059] The specific execution process of step S102 is as follows: first, obtain the industrial and commercial information of the sample enterprises, as well as the number of researchers and the proportion of academic qualifications, qualification certification information, the number of Class I intellectual property rights, the number of Class II intellectual property rights, project application information, loan amount, financing data, R&D expenses and technical contract information in each of the past three years, among which the loan amount includes the maximum loan amount, the minimum loan amount, the average loan amount and the total loan amount; then, clean and normalize the content of the above-mentioned acquired data, replace the Chinese content with digital representation, and obtain numerical enterprise data; secondly, label and classify the numerical enterprise data to determine each indicator item, put the data of each indicator item into the object of the enterprise data matrix, and calculate the data value of the sample enterprise under each indicator item respectively; finally, calculate the sum, average, maximum, minimum, 25% and 75% values ​​of the data values ​​of all sample enterprises under each indicator item, and process abnormal data values.

[0060] The index items of the above sample enterprises include the total number of employees, the proportion of researchers, the proportion of college degree and above, the proportion of postgraduates, the number of enterprise qualifications, industry-university-research qualifications, the number of subordinate institutions, the number of R&D activities, the proportion of R&D expenses, the enterprise's R&D expenses in the previous year, the R&D expenses last year, the project establishment expenses in the previous year, the project establishment expenses last year, the maximum project amount, the minimum project amount, the average project amount, the total project amount, the total number of approved projects, the total number of applied projects, the total amount of equity investment, the total number of equity investments, the total amount of debt in the previous year, the total amount of debt last year, the maximum amount of debt, the minimum amount of debt, the average amount of debt, the total amount of debt, the total number of debts obtained, the total assets in the previous year, the total assets last year, the net assets in the previous year, the net assets last year, the main business income in the previous year, the main business income last year, the cost and expenses in the previous year, the cost and expenses last year, the net profit in the previous year, the net profit last year, the previous year Profits and taxes paid, profits and taxes paid last year, proportion of main business in the previous year, proportion of main business in last year, net profit margin in the previous year, net profit margin last year, management certification, Class I intellectual property rights, Class II intellectual property rights, number of technology contracts in the previous year, number of technology contracts last year, maximum technology transaction amount of technology contracts, minimum technology transaction amount of technology contracts, average technology transaction amount of technology contracts, total technology transaction amount of technology contracts, total number of technology transaction amounts of technology contracts, number of national science and technology awards, number of standards formed, number of scientific and technological achievements transformed, number of high-tech products, high-tech income, growth rate of total assets in the previous year, growth rate of net assets, growth rate of main business income, rate of change of cost and expenses, net profit growth rate, growth rate of profits and taxes paid, growth rate of R&D expenses, rate of change of proportion of main business, growth rate of net profit margin, growth rate of proportion of R&D expenses, growth rate of number of intellectual property rights and growth rate of proportion of R&D personnel.

[0061] The process of acquiring, cleaning, calculating and classifying sample data of each dimension of sample enterprises at least includes the following: Figures 2 to 4 The 3 stages are shown.

[0062] like Figure 2As shown, in the early stage, the data preparation stage is entered to process the operating income data, and the relevant data required for the evaluation index model of all sample enterprises are exported from the database. The relevant data covers at least the above-mentioned 8 enterprise dimensions. The relevant data forms the "All Enterprises.xls" document and is stored in the memory for backup; a set of all non-repeating sample enterprise names is obtained from the "All Enterprises.xls" document, and for any sample enterprise in the set: first obtain the latest science and technology innovation loan data of the enterprise as the data basis of the enterprise in this evaluation method; then calculate the maximum, minimum, average and total science and technology innovation loan amounts obtained by the enterprise in the past three years, 2) the average proportion of researchers in each year, 3) the average number of Class I intellectual property rights in each year, and 4) the average number of Class II intellectual property rights in each year to form a data object CA; for any sample enterprise in the set, the latest science and technology innovation loan data of the enterprise and its corresponding data object CA are merged to form new enterprise data and transferred to the initial table "date1.xls" file in the first step.

[0063] Normalize the data content in the initial table of the first step. For example, remove the Chinese characters in the columns of the proportion of researchers, the proportion of college education and above, and the total number of employees in the table, and convert them into float type data; convert the no / yes in the table indicating whether it is a high-tech enterprise, a technologically advanced enterprise, or a technology-based small and medium-sized enterprise to 0 / 1; and convert the enterprise data type of the specified column in the initial table of the first step into float type data, and merge the enterprise information of each sample enterprise and the corresponding enterprise data of the enterprise into the first step result table "sheet1.xlsx" file.

[0064] like Figure 3As shown in the figure, the mid-term quantitative evaluation of the sample enterprises' ability to convert achievements is carried out by label classification, which includes obtaining five parts of data; the first is the result table "sheet1.xlsx" file of the first step; the second is the enterprise data on several qualification certification information and the enterprise's intellectual property and researcher data obtained from the initial table "date1.xls" file of the first step; for qualification certification, all enterprises' non-repeated certification types are sorted according to the separators in the table, and converted into type columns according to the certification type. If the column value of the enterprise is thus certified, it is recorded as 1; the third is to obtain the equity financing data of all enterprises from the database, take the intersection of this data with all relevant data of the above-mentioned "all enterprises.xls" document, obtain the first intersection data, and calculate any sample enterprise according to the first intersection data. Calculate the number of investments and the total investment amount obtained by the enterprise; fourthly, obtain the technology contract data of all enterprises from the database, take the intersection of this data with all relevant data in the above-mentioned "all enterprises.xls" document to obtain the second intersection data, and calculate the technology transaction amount of the enterprise in each of the past three years and the maximum, minimum, average, total data and number of technology transaction amounts of the enterprise for any sample enterprise according to the second intersection data; fifthly, obtain the project application data of all enterprises from the database, take the intersection of this data with all relevant data in the above-mentioned "all enterprises.xls" document to obtain the third intersection data, and calculate the project approval amount of the enterprise in each of the past three years and the maximum, minimum, average, total data and number of project approval amounts of the enterprise for any sample enterprise according to the third intersection data.

[0065] Finally, the above five parts of data are merged according to the above-mentioned enterprise dimensions, and the second step result table "date99.xlsx" file is output.

[0066] like Figure 4As shown in the figure, the data values ​​and abnormal data values ​​of the sample enterprises under each indicator are calculated in the later stage. In this stage, the data of the early and mid-term are combined; that is, the operating income in the first step result table "date1.xls" file is merged with the second step result table "date99.xlsx" file, and the total amount of subsidies obtained by each sample enterprise (industry-university-research), the number of qualification certifications obtained by each sample enterprise (enterprise qualification), the number of management certifications obtained by each enterprise (management certification), the asset, profit, income, tax payment and R&D expense growth rate of the sample enterprise, and the number of personnel of the sample enterprise are calculated in turn. The calculated data are then merged with the data in the second step result table "date99.xlsx" file and recorded as the indicator item data "bb_date" table object for all sample companies; for each column of the indicator item data "bb_date" table object, the count, mean, std, min, 25%, 50%, 75%, max, null_% (the proportion of the number of columns and rows with data to the total number of rows) of the column data are calculated respectively, and the calculated values ​​are recorded as the "n_df" table object.

[0067] Finally, the abnormal data values ​​in the indicator item data "bb_date" table object are processed. The processing method of the abnormal data values ​​proposed in the embodiment is: for data values ​​whose proportion data are not in the range of 0 to 1, when it is greater than 1, it is set to 75% of the corresponding item, and when it is less than 1, it is set to 25% of the corresponding item; for data with a growth rate greater than 20, it is set to 75% of the corresponding item, and for data with a growth rate less than 20, it is set to 25% of the corresponding item; the values ​​of some label columns that exceed the maximum value are set to the corresponding maximum value, including the total number of approved projects, the number of subordinate institutions, the total number of applied projects, the maximum amount of debt, the average amount of debt, and the total amount of debt; after the abnormal data values ​​are processed, the processed data is stored in the third step result data "data_72.xlsx" file.

[0068] Step S104, using the sample data and label classification after cleaning of the sample enterprises to train the indicator weight model, so as to obtain the relative importance scores and priority diagrams of the indicator items according to the indicator weight model, and then calculate the final weight value of each indicator item;

[0069] The specific execution process of step S104 is as follows: predefine a multi-level data classification label as a calculation index item; receive all sample data of the sample enterprise, use the gradient boosting decision tree algorithm of the machine learning xgboost library to extract the characteristics of each sample data and classify, sort and group them, combine the input calculation index items to perform softmax loss function plus regularization term calculation, and obtain the relative importance score and priority diagram of each index item one by one; take the average of the relative importance score of the index item and the preset corresponding index item proportion score and the final weight value of the index item.

[0070] Figure 5 The calculation process of the final weight value with the total amount of the project amount as the calculation indicator item is shown. Specifically, first, the total amount of the project amount column in the third step result data "data_72.xlsx" is copied to column Y, and the data in column Y is preprocessed, including excluding all rows with null values ​​in column Y and retaining rows with at least 50 non-null columns; secondly, the gradient boosting decision tree algorithm of the xgboost library is used to train the weight model, the feature matrix is ​​the data outside column Y, and the label is column Y; then, the importance score result of the indicator item is obtained, and the result is bound to the column name of the feature matrix source, that is, the binding result is the column name of the third step result data "data_72.xlsx", and the value is the importance score; then, based on the binding result, the feature matrix source column index number is added to form number column 1, and number column 1 is sorted in ascending order according to the importance score; the feature matrix source row index number is added to number column 1 to form number column 2, and the column name and value of number column 2 are obtained according to the sorting of number column 1.

[0071] Finally, merge the numbered column 2 and the "comprehensive weight" and "index" column data read from the preset all indicator weight distribution table "zonghe.xlsx" to calculate the final weight value of the indicator item. The calculation method is: (the value of the numbered column 2 + comprehensive weight) / 2. Calculate the indicator items of the enterprise dimension in turn, and save all the results as the weight data "qz.xlsx" file.

[0072] Step S106, calculate the final evaluation score and ranking of each indicator item of each enterprise to be evaluated based on the enterprise data of each enterprise to be evaluated and the final weight value of each indicator item; wherein the enterprise data is the sample data of each enterprise dimension after data cleaning of the enterprise to be evaluated.

[0073] The specific execution process of step S108 is: obtaining sample data of each enterprise dimension of each enterprise to be evaluated and performing data processing to obtain the enterprise data matrix of each enterprise to be evaluated; dividing the indicator items calculated by the indicator weight model into five groups of evaluation dimensions, including endogenous innovation, collaborative investment, management and operation capabilities, results transformation capabilities and sustainable growth capabilities, and performing matrix dot multiplication on the corresponding numerical value of any indicator item in each group of evaluation dimensions in the enterprise data matrix and its corresponding final weight value to obtain the evaluation score of the corresponding indicator item; according to the normal distribution characteristics, converting the scores of each group of evaluation dimensions into a 100-point system as the final evaluation score of the enterprise indicator item.

[0074] An embodiment of the present application also provides an electronic device, which includes a processor and a memory, the memory being used to store program instructions and transmit the program instructions to the processor; when the program instructions are executed by the processor, the processor runs the evaluation method for enterprise innovation and growth capabilities based on big data analysis in the above embodiment.

[0075] The above program can be run in a processor, or can also be stored in a computer-readable storage medium, which includes permanent and non-permanent, removable and non-removable media. Information storage can be achieved by any method or technology. Information can be computer-readable instructions, data structures, modules of programs or other data. Examples of computer-readable storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, read-only compact disk read-only memory (CD-ROM), digital versatile disk (DVD) or other optical storage, magnetic cassettes, magnetic tape magnetic disk storage or other magnetic storage devices or any other non-transmission media that can be used to store information that can be accessed by a computing device. As defined herein, computer-readable storage media does not include temporary computer-readable media (transitory media), such as modulated data signals and carrier waves.

[0076] These computer programs can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, so that the instructions executed on the computer or other programmable device provide instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps of the functions specified in one or more blocks can be implemented by different modules corresponding to different steps.

[0077] Exemplarily, this embodiment provides such a device or system, namely, an evaluation system for enterprise innovation and growth capabilities based on big data analysis, which includes the following program modules: an enterprise data collection and processing module, which is used to collect sample data of each enterprise dimension of sample enterprises, and to clean, calculate and label the sample data according to time series; wherein the enterprise dimensions include industry and commerce, employees, qualifications, intellectual property rights, projects, financing loans, assets and R&D expenses; an indicator weight module, which is used to train the indicator weight model according to the cleaned sample data and label classification of the sample enterprises, so as to obtain the relative importance score and priority diagram of the indicator items according to the indicator weight model, and then calculate the final weight value of each indicator item; a score assignment module, which is used to calculate the final evaluation score and ranking of each indicator item of each enterprise to be evaluated based on the enterprise data of each enterprise to be evaluated in combination with the final weight value of each indicator item.

[0078] The system is used to implement the functions of the evaluation method in the above-mentioned embodiment. Each module in the system corresponds to each step in the method, which has been explained in the method and will not be repeated here.

[0079] For example, the enterprise data collection and processing module includes: a first acquisition unit, which is used to obtain the industrial and commercial information of the sample enterprise, as well as the number of researchers and the proportion of academic qualifications, qualification certification information, the number of Class I intellectual property rights, the number of Class II intellectual property rights, project application information, loan amount, financing data, R&D expenses and technical contract information in each of the past three years, among which the loan amount includes the maximum loan amount, the minimum loan amount, the average loan amount and the total loan amount; a cleaning and normalization unit, which is used to clean and normalize the data obtained by the acquisition unit, including deleting duplicate data, evaluating irrelevant data and replacing Chinese content with digital representation to obtain numerical enterprise data; a classification unit, which is used to label and classify the numerical enterprise data to determine each indicator item, put the data of each indicator item into the object of the enterprise data matrix, and calculate the data value of the sample enterprise under each indicator item respectively; a calculation unit, which is used to calculate the sum, average, maximum, minimum, 25% and 75% values ​​of the data values ​​of all sample enterprises under each indicator item, and process abnormal data values.

[0080] For another example, the indicator weight module includes: a predefined unit, which is used to predefine a multi-level data classification label as a calculation indicator item; a machine learning unit, which is used to receive all sample data of the sample enterprise, use the gradient boosting decision tree algorithm of the machine learning xgboost library to extract the characteristics of each sample data and classify, sort and group them, and combine the input calculation indicator items to perform softmax loss function plus regularization term calculation, and obtain the relative importance score and priority diagram of each indicator item one by one; a weight calculation unit, which is used to calculate the relative importance score of the indicator item and the mean of the sum of the preset corresponding indicator item proportion scores, and use the mean as the final weight value of the indicator item.

[0081] And for example, the score assignment module includes: a second acquisition unit, used to obtain sample data of each enterprise dimension of each enterprise to be evaluated and perform data processing to obtain the enterprise data matrix of each enterprise to be evaluated; a division evaluation unit, used to divide the various indicator items calculated by the indicator weight model into five groups of evaluation dimensions, including endogenous innovation, collaborative investment, management and operation capabilities, results transformation capabilities and sustainable growth capabilities, and perform matrix dot multiplication on the corresponding numerical value of any indicator item in each group of evaluation dimensions in the enterprise data matrix and its corresponding final weight value to obtain the evaluation score of the corresponding indicator item; a score conversion unit, used to convert the scores of each group of evaluation dimensions into a 100-point system according to the normal distribution characteristics as the final evaluation score of the enterprise indicator item.

[0082] The embodiments of the present application solve the problem that financial analysis methods are not suitable for evaluating the growth capabilities of entrepreneurial technology companies with a short establishment time. By quantitatively evaluating the non-quantifiable innovation capabilities and results transformation capabilities in the business activities of the company, it avoids focusing on the company's financial indicators. At the same time, it is suitable for the growth capability evaluation of large, medium, small and micro enterprises, providing more accurate and reliable data support for the company's development planning, policy matching and corporate investment and financing decisions.

[0083] Although the present invention has been disclosed as above with preferred embodiments, it is not intended to limit the present invention. A person with ordinary knowledge in the technical field to which the present invention belongs may make various changes and modifications without departing from the spirit and scope of the present invention. Therefore, the protection scope of the present invention shall be determined by the definition of the claims.

Claims

1. A method for evaluating enterprise innovation and growth capabilities based on big data analysis, characterized in that: include: Collect sample data of each enterprise dimension of sample enterprises, and clean, calculate and label the sample data according to time series; the enterprise dimensions include industry and commerce, employees, qualifications, intellectual property rights, projects, financing loans, assets and R&D expenses; The cleaned sample data and label classification of sample enterprises are used to train the indicator weight model, so as to obtain the relative importance scores and priority diagrams of the indicator items according to the indicator weight model, and then calculate the final weight value of each indicator item; The final evaluation score and ranking of each indicator item of each enterprise to be evaluated are calculated based on the enterprise data of each enterprise to be evaluated and the final weight value of each indicator item; among which, the enterprise data is the sample data of each enterprise dimension after data cleaning of the enterprise to be evaluated; The calculation process of the final weight value of the indicator item is: Predefine a multi-level data classification label as a calculation indicator item; receive all sample data of sample enterprises, use the gradient boosting decision tree algorithm of the machine learning xgboost library to extract the features of each sample data and classify, sort and group them, combine the input calculation indicator items to perform softmax loss function plus regularization term calculation, and obtain the relative importance score and priority diagram of each indicator item one by one; take the average of the relative importance score of the indicator item and the preset corresponding indicator item proportion score and the final weight value of the indicator item; The calculation process of the final evaluation score of each indicator item of each enterprise to be evaluated is as follows: Obtain sample data of each enterprise dimension of each enterprise to be evaluated and perform data processing to obtain the enterprise data matrix of each enterprise to be evaluated; divide the indicator items calculated by the indicator weight model into five groups of evaluation dimensions, including endogenous innovation, collaborative investment, management and operation capabilities, results transformation capabilities and sustainable growth capabilities, and perform matrix dot multiplication on the corresponding numerical value of any indicator item in each group of evaluation dimensions in the enterprise data matrix and its corresponding final weight value to obtain the evaluation score of the corresponding indicator item; according to the normal distribution characteristics, convert the scores of each group of evaluation dimensions into a 100-point system as the final evaluation score of the enterprise indicator item.

2. The evaluation method for enterprise innovation and growth capability based on big data analysis according to claim 1 is characterized in that: The process of data cleaning, calculation and label classification of sample data according to time series includes: Obtain the business information of the sample enterprises, as well as the number of researchers and their educational background, qualification certification information, the number of Class I intellectual property rights, the number of Class II intellectual property rights, project application information, loan amount, financing data, R&D expenses and technical contract information in each of the past three years. The loan amount includes the maximum loan amount, the minimum loan amount, the average loan amount and the total loan amount. Clean and normalize the above-obtained data, replace the Chinese content with digital representation, and obtain numerical enterprise data; Label and classify the numerical enterprise data to determine the indicators, put the data of each indicator into the object of the enterprise data matrix, and calculate the data value of the sample enterprise under each indicator; Calculate the sum, average, maximum, minimum, 25% and 75% values ​​of the data values ​​of all sample enterprises under each indicator, and handle abnormal data values.

3. The evaluation method for enterprise innovation and growth capability based on big data analysis according to claim 2 is characterized in that: The index items of the sample enterprises include the total number of employees, the proportion of researchers, the proportion of college graduates and above, the proportion of postgraduates, the number of enterprise qualifications, industry-university-research qualifications, the number of subordinate institutions, the number of R&D activities, the proportion of R&D expenses, the enterprise's R&D expenses in the previous year, the R&D expenses last year, the project establishment expenses in the previous year, the project establishment expenses last year, the maximum project amount, the minimum project amount, the average project amount, the total project amount, the total number of approved projects, the total number of project applications, the total amount of equity investment, the total number of equity investments, the total amount of debt in the previous year, the total amount of debt last year, the maximum amount of debt, the minimum amount of debt, the average amount of debt, the total amount of debt, the total number of debts obtained, the total assets in the previous year, the total assets last year, the net assets in the previous year, the net assets last year, the main business income in the previous year, the main business income last year, the cost and expenses in the previous year, the cost and expenses last year, the net profit in the previous year, the net profit last year, the previous year Profits and taxes paid, profits and taxes paid last year, proportion of main business in the previous year, proportion of main business in last year, net profit margin in the previous year, net profit margin last year, management certification, Class I intellectual property rights, Class II intellectual property rights, number of technology contracts in the previous year, number of technology contracts last year, maximum technology transaction amount of technology contracts, minimum technology transaction amount of technology contracts, average technology transaction amount of technology contracts, total technology transaction amount of technology contracts, total number of technology transaction amounts of technology contracts, number of national science and technology awards, number of standards formed, number of scientific and technological achievements transformed, number of high-tech products, high-tech income, growth rate of total assets in the previous year, growth rate of net assets, growth rate of main business income, rate of change of cost and expenses, net profit growth rate, growth rate of profits and taxes paid, growth rate of R&D expenses, rate of change of proportion of main business, growth rate of net profit margin, growth rate of proportion of R&D expenses, growth rate of number of intellectual property rights and growth rate of proportion of R&D personnel.

4. The method for evaluating enterprise innovation and growth capabilities based on big data analysis according to claim 2 is characterized in that: The method for processing abnormal data values ​​is as follows: for data values ​​whose proportion data are not in the range of 0 to 1, when they are greater than 1, they are set to 75% of the corresponding item, and when they are less than 1, they are set to 25% of the corresponding item; for data with a growth rate greater than 20, they are set to 75% of the corresponding item, and data with a growth rate less than 20 are set to 25% of the corresponding item.

5. An evaluation system for enterprise innovation and growth capabilities based on big data analysis, characterized in that: include: The enterprise data collection and processing module is used to collect sample data of various enterprise dimensions of sample enterprises, and clean, calculate and label the sample data according to the time series; the enterprise dimensions include industry and commerce, employees, qualifications, intellectual property rights, projects, financing loans, assets and R&D expenses; The indicator weight module is used to train the indicator weight model based on the sample data and label classification after cleaning of the sample enterprises, so as to obtain the relative importance scores and priority diagrams of the indicator items according to the indicator weight model, and then calculate the final weight value of each indicator item; The score assignment module is used to calculate the final evaluation score and ranking of each indicator item of each enterprise to be evaluated based on the enterprise data of each enterprise to be evaluated and the final weight value of each indicator item; The indicator weight module includes: The predefined unit is used to predefine a multi-level data classification label as a calculation indicator item; the machine learning unit is used to receive all sample data of the sample enterprise, use the gradient boosting decision tree algorithm of the machine learning xgboost library to extract the characteristics of each sample data and classify, sort and group, and perform softmax loss function plus regularization calculation on the input calculation indicator items to obtain the relative importance score and priority diagram of each indicator item one by one; the weight calculation unit is used to calculate the relative importance score of the indicator item and the mean of the sum of the preset corresponding indicator item proportion scores, and use the mean as the final weight value of the indicator item; The score assignment module includes: The second acquisition unit is used to obtain sample data of each enterprise dimension of each enterprise to be evaluated and perform data processing to obtain the enterprise data matrix of each enterprise to be evaluated; the division evaluation unit is used to divide the various indicator items calculated by the indicator weight model into five groups of evaluation dimensions, including endogenous innovation, collaborative investment, management and operation capabilities, results transformation capabilities and sustainable growth capabilities, and perform matrix dot multiplication on the corresponding numerical value of any indicator item in each group of evaluation dimensions in the enterprise data matrix and its corresponding final weight value to obtain the evaluation score of the corresponding indicator item; the score conversion unit is used to convert the scores of each group of evaluation dimensions into a 100-point system according to the normal distribution characteristics as the final evaluation score of the enterprise indicator item.

6. The evaluation system for enterprise innovation and growth capabilities based on big data analysis according to claim 5 is characterized in that: The enterprise data collection and processing module includes: The first acquisition unit is used to obtain the industrial and commercial information of the sample enterprises, as well as the number of researchers and the educational background ratio, qualification certification information, the number of Class I intellectual property rights, the number of Class II intellectual property rights, project application information, loan amount, financing data, R&D expenses and technical contract information in each of the past three years. The loan amount includes the maximum loan amount, the minimum loan amount, the average loan amount and the total loan amount. A cleaning and normalization unit is used to clean and normalize the data acquired by the acquisition unit, including deleting duplicate data, evaluating irrelevant data, and replacing Chinese content with digital representation to obtain numerical enterprise data; The classification unit is used to classify the numerical enterprise data into labels to determine each indicator item, put the data of each indicator item into the object of the enterprise data matrix, and calculate the data value of the sample enterprise under each indicator item respectively; The calculation unit is used to calculate the sum, average, maximum, minimum, 25% and 75% values ​​of the data values ​​of all sample enterprises under each indicator item, and process abnormal data values.

Images