Research and development efficiency processing method and device, terminal and storage medium

Abstract

The embodiment of the invention provides a research and development efficiency processing method and device, a terminal and a storage medium. The research and development efficiency processing method comprises the steps of determining at least one to-be-evaluated product department; selecting a second index from the first indexes based on the first data of the first indexes of at least one product department; inputting second data of a second index of the at least one product department into the efficiency evaluation model to obtain an efficiency result of the at least one product department; the efficiency result is used for indicating the efficiency sequence of at least one product department; constructing a linear regression model by taking the efficiency result as a dependent variable and taking second input index data of the second input index as an independent variable; and based on the linear regression model, carrying out significance test on the independent variables to obtain the independent variables with significance parameters of significance test higher than a threshold value as a third input index.

Description

Technical Field

[0001] The present invention relates to, but is not limited to, the field of data processing technologies, and particularly relates to a research and development (R&D) efficiency processing method, apparatus, terminal, and storage medium. Background Art

[0002] With the continuous improvement of the digitalization level of enterprises, R&D efficiency management is becoming a key concern for enterprises. Only under efficient and orderly R&D management can an enterprise continuously innovate, enhance its core competitiveness, reduce costs, and gain a competitive advantage in the market. Having efficient and accurate R&D efficiency evaluation and / or analysis capabilities can help enterprises effectively improve organizational efficiency. Among them, the goal of R&D efficiency evaluation and analysis of factors affecting R&D efficiency is to achieve quantitative evaluation of enterprise R&D efficiency indicators and instantaneously identify bottlenecks and improvement points in the R&D process by constructing a reasonable R&D efficiency evaluation index system, so as to continuously improve the enterprise's R&D efficiency and product competitiveness.

[0003] However, there are still many problems in the current processing of R&D efficiency evaluation and / or analysis; for example, there is a large degree of human uncertainty, and it is impossible to intuitively compare the R&D efficiency levels of different departments within an enterprise. Another example is that it is impossible to conduct statistical analysis on the factors affecting the enterprise's R&D efficiency, so it is impossible to provide effective suggestions for improving and optimizing the R&D efficiency. Summary of the Invention

[0004] In view of this, embodiments of the present invention provide a R&D efficiency processing method, apparatus, terminal, and storage medium to solve the above technical problems.

[0005] The technical solution of the present invention is realized as follows:

[0006] In a first aspect, an embodiment of the present invention provides a R&D efficiency processing method, the method including:

[0007] Determine at least one product department to be evaluated;

[0008] Based on the first data of the first indicators of at least one product department, select second indicators from the first indicators; wherein, the first indicators include first input indicators and first output indicators; the first data includes first input indicator data of the first input indicators and first output indicator data of the first output indicators; the second indicators include second input indicators selected from the first input indicators and second output indicators selected from the first output indicators;

[0009] Input the second data of the second indicators of at least one product department into an efficiency evaluation model to obtain the efficiency results of at least one product department; wherein, the second data includes second input indicator data of the second input indicators and second output indicator data of the second output indicators; the efficiency results are used to indicate the efficiency ranking of at least one product department.

[0010] Taking the performance result as the dependent variable and the second input index data of the second input index as the independent variable to construct a linear regression model;

[0011] Based on the linear regression model, perform a significance test on the independent variable, and use the independent variable with a significance parameter higher than the threshold obtained from the significance test as the third input index.

[0012] In some embodiments, selecting a second index from the first index based on the first data of the first index of at least one product department includes:

[0013] Performing a distinguishability test on the first data to obtain the first index with a distinguishability greater than or equal to a predetermined distinguishability as the first candidate index; wherein, the first candidate index includes a first candidate input index and a first candidate output index;

[0014] Performing a redundancy test on the first candidate index to obtain the first candidate index with a correlation coefficient less than or equal to a predetermined correlation coefficient as the second index; wherein, the correlation coefficient is used to characterize the correlation between the first candidate indexes.

[0015] In some embodiments, performing a distinguishability test on the first data to obtain the first index with a distinguishability greater than or equal to a predetermined distinguishability as the first candidate index includes:

[0016] Determining the standard deviation of the first data of the first index of each product department; wherein, the number of the first indexes is the number of the standard deviations, the average value of the first data of each first index is the average value of the standard deviations, and the first data of each first index is the data point of the standard deviation;

[0017] Determining the standard deviation as the distinguishability for performing the distinguishability test on the first data;

[0018] If the distinguishability of the product department is greater than or equal to the predetermined distinguishability, determining the first index of the product department as the first candidate index.

[0019] In some embodiments, the performance evaluation model includes a first performance evaluation model and a second performance evaluation model. The first performance evaluation model is a non - linear model, and the second performance evaluation model is a non - linear evaluation model; inputting the second data of the second index of at least one product department into the performance evaluation model to obtain the performance result of at least one product department includes:

[0020] Inputting the number of the second input indexes, the second input index data of each second input index, the number of the second output indexes, and the second output index data of each second output index into the first performance evaluation model;

[0021] Converting the first performance evaluation model into the second performance evaluation model;

[0022] Based on the second performance evaluation model, to obtain the performance results of at least one product department.

[0023] In some embodiments, the method includes:

[0024] Determine a window period for the second indicators of at least one product department;

[0025] Determine the second indicators within a window period as second candidate indicators, where the second candidate indicators include second candidate input indicators corresponding to the second input indicators and second candidate output indicators corresponding to the second output indicators;

[0026] Input the quantity of the second input indicators, the second input indicator data of each second input indicator, the quantity of the second output indicators, and the second output indicator data of each second output indicator into the first performance evaluation model, including: inputting the quantity of the second candidate input indicators, the second candidate input indicator data of each second candidate input indicator, the quantity of the second candidate output indicators, and the second candidate output indicator data of each second candidate output indicator into the first performance evaluation model.

[0027] In some embodiments, before constructing the linear regression model, it includes:

[0028] Perform at least one of the following tests on the second input indicator data of the second input indicators to obtain the second input indicators that pass the tests: stationarity test, Granger causality test, and multicollinearity test;

[0029] Use the performance results as the dependent variable and the second input indicator data of the second input indicators as the independent variables to construct a linear regression model, including:

[0030] Use the performance results as the dependent variable and the second input indicator data of the second input indicators that pass the tests as the independent variables to construct a linear regression model.

[0031] In some embodiments, the method further includes:

[0032] Perform at least one of the analysis of the regression coefficient, positive / negative nature, P-value, and significance of the third input indicators of at least one product department to obtain the influence results of each third input indicator on at least one product department, where the influence results characterize the magnitude of the influence degree and / or the magnitude of the influence correlation;

[0033] The regression coefficient is a parameter indicating the magnitude of the influence degree of the independent variable on the dependent variable;

[0034] The positive / negative nature includes positive correlation or negative correlation. Positive correlation is a parameter indicating the magnitude of the increase of the dependent variable as the independent variable increases, and negative correlation is a parameter indicating the magnitude of the decrease of the dependent variable as the independent variable increases;

[0035] The P - value is used to indicate the significance of the regression coefficient;

[0036] Significance is used to indicate the difference between the regression coefficient and zero; Significance is represented by the P - value.

[0037] In a second aspect, an R & D efficiency processing device according to an embodiment of the present invention includes:

[0038] A determination module, configured to determine at least one product department to be evaluated;

[0039] A first processing module, configured to select a second indicator from the first indicators based on the first data of the first indicators of at least one product department; wherein, the first indicators include a first input indicator and a first output indicator; the first data includes first input indicator data of the first input indicator and first output indicator data of the first output indicator; the second indicators include a second input indicator selected from the first input indicators and a second output indicator selected from the first output indicators;

[0040] A second processing module, configured to input the second data of the second indicators of at least one product department into an efficiency evaluation model to obtain an efficiency result of at least one product department; wherein, the second data includes second input indicator data of the second input indicator and second output indicator data of the second output indicator; the efficiency result is used to indicate the efficiency ranking of at least one product department;

[0041] The second processing module, configured to use the efficiency result as the dependent variable and the second input indicator data of the second input indicator as the independent variable to construct a linear regression model;

[0042] The second processing module, configured to perform a significance test on the independent variable based on the linear regression model to obtain the independent variable with a significance parameter of the significance test higher than the threshold as the third input indicator.

[0043] In a third aspect, an embodiment of the present invention provides a terminal, which includes a processor and a memory for storing a computer program that can run on the processor; wherein, when the processor runs the computer program, it implements the R & D efficiency processing method of any embodiment of the present invention.

[0044] In a fourth aspect, an embodiment of the present invention further provides a computer storage medium, in which there are computer - executable instructions, and when the computer - executable instructions are executed by a processor, the R & D efficiency processing method of any embodiment of the present invention is implemented.

[0045] In a fifth aspect, an embodiment of the present invention provides a computer program product, which includes a computer program or instructions, and when the computer program or instructions are executed by a processor, the R & D efficiency processing method of any embodiment of the present invention is implemented.

[0046] In an embodiment of the present invention, a linear regression model can be constructed by taking the effectiveness result as the dependent variable and the second input index as the independent variable, so as to effectively analyze the influencing factors of the R & D effectiveness of at least one department's products based on the linear regression model, and thus obtain the third input index with significant effectiveness influence. In this way, on the one hand, the evaluation of efficiency in the R & D process is considered, which is convenient for intuitively evaluating the R & D effectiveness of multiple product departments within an enterprise. On the other hand, the input index can be calculated through the linear regression model, reducing the influence of subjective factors in the effectiveness evaluation and improving the accuracy of the effectiveness evaluation; moreover, it can also conduct statistical analysis on the influencing factors of the R & D effectiveness, facilitating the provision of effective suggestions for subsequent improvement and optimization of the R & D effectiveness. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] Figure 1 It is a schematic flowchart of the first R & D effectiveness processing method provided by an embodiment of the present invention.

[0048] Figure 2 It is a schematic flowchart of the second R & D effectiveness processing method provided by an embodiment of the present invention.

[0049] Figure 3 It is a schematic flowchart of the third R & D effectiveness processing method provided by an embodiment of the present invention.

[0050] Figure 4 It is a schematic flowchart of the fourth R & D effectiveness processing method provided by an embodiment of the present invention.

[0051] Figure 5 It is a schematic structural diagram of an R & D effectiveness processing device provided by an embodiment of the present invention.

[0052] Figure 6 It is a schematic hardware structure diagram of a terminal provided by an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0053] The present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0054] In the following description, suffixes such as "module", "component", or "unit" used to represent elements are only for facilitating the description of the present invention and have no specific meaning in themselves. Therefore, "module", "component", or "unit" can be used interchangeably. Additionally, in the subsequent statements, prefixes such as "first" or "second" used to identify information are only for facilitating the description of the present invention and have no specific meaning in themselves. Further, in the following description, "a plurality" means two or more; "a variety" means two or more.

[0055] As Figure 1 shown, an embodiment of the present invention provides a research and development efficiency, including the following steps:

[0056] Step S101: Determine at least one product department to be evaluated;

[0057] Step S102: Select a second indicator from the first indicators based on the first data of the first indicators of at least one product department; wherein, the first indicators include a first input indicator and a first output indicator; the first data includes the first input indicator data of the first input indicator and the first output indicator data of the first output indicator; the second indicators include a second input indicator selected from the first input indicators and a second output indicator selected from the first output indicators;

[0058] Step S103: Input the second data of the second indicators of at least one product department into the efficiency evaluation model to obtain the efficiency results of at least one product department; wherein, the second data includes the second input indicator data of the second input indicator and the second output indicator data of the second output indicator; the efficiency results are used to indicate the efficiency ranking of at least one product department;

[0059] Step S104: Use the efficiency results as the dependent variable and the second input indicator data of the second input indicator as the independent variable to construct a linear regression model;

[0060] Step S105: Based on the linear regression model, perform a significance test on the independent variable to obtain the independent variable with a significance parameter of the significance test higher than the threshold as the third input indicator.

[0061] The research and development efficiency processing method provided by the embodiment of the present invention is executed by a terminal; the terminal can be any kind of mobile terminal or fixed terminal. For example, the terminal can be, but is not limited to, at least one of the following: a mobile communication device, a computer, a server, a tablet computer, a smart office device, an industrial device, and / or a wearable device, etc. The mobile communication device can be, but is not limited to, a mobile phone or a smart phone.

[0062] Optionally, at least one product department may be one or more product departments. In an embodiment of the present invention, at least one means one or more; and more than one means two or more.

[0063] Optionally, in an embodiment of the present invention, a product department may be a product-related department of any enterprise, company, unit, organization or institution. For example, the product department may be, but is not limited to, at least one of the following: a research and development department, a technology department, a production department, a testing department, a product management department, a quality department, a sales department, a quality control department, and a research department.

[0064] Optionally, the first indicator, the second indicator, the third indicator, and the first candidate indicator, the second candidate indicator, the first preliminary selection indicator, and the second preliminary selection indicator involved below may all include input indicators and / or output indicators.

[0065] Exemplarily, the first indicator may include a first input indicator and / or a first output indicator; the second indicator may include a second input indicator and / or a second output indicator; the third indicator may include a third input indicator and / or a third output indicator.

[0066] Exemplarily, the first candidate indicator may include a first candidate input indicator and / or a first candidate output indicator; the second candidate indicator may include a second candidate input indicator and / or a second candidate output indicator.

[0067] Exemplarily, the input indicators may include, but are not limited to, at least one of the following: a first input indicator, a second input indicator, a third input indicator, a first candidate input indicator, and a second candidate input indicator, etc.

[0068] Exemplarily, the output indicators may include, but are not limited to, at least one of the following: a first output indicator, a second output indicator, a third output indicator, a first candidate output indicator, and a second candidate output indicator, etc.

[0069] Exemplarily, the input indicator may be an indicator related to product input. For example, the input indicator may include, but is not limited to, at least one of the following: research and development man-hours, quality inspection man-hours, raw materials, funds, externally purchased technologies, and the job levels of employees, etc. Externally purchased technologies refer to resources, information, and / or data collected or purchased from outside; for example, externally purchased technologies may be codes and / or products, etc. The research and development man-hours may include, but are not limited to, the research and development man-hours for each product and / or itemized reported man-hours, etc. The job levels may include, but are not limited to, senior, intermediate, or junior levels, etc.; or the job levels may include any level from the first level to the tenth level, etc.

[0070] Exemplarily, the output metrics may be metrics related to product output. For example, the output metrics may include, but are not limited to, at least one of the following: mileage of the product, quantity of the product, revenue of the product, result indicators of the product, papers and / or patents related to the product, etc. The mileage of the product may refer to different stages or milestones experienced by the product in its life cycle.

[0071] Optionally, both the input metrics and the output metrics may be divided into first-level or multi-level metrics.

[0072] Exemplarily, the input metrics may be divided into two-level input metrics; for example, the input metrics may include first-level input metrics and second-level input metrics; the first-level input metrics may include at least one second-level input metric. For example, the first-level input metric is R & D man-hours, and the second-level input metrics may include R & D man-hours for the product and / or itemized reporting man-hours. Of course, the input metrics may also be divided into three levels or more levels of input metrics; for example, when the input metrics are divided into three-level input metrics, they may include first-level input metrics, second-level input metrics, and third-level input metrics, where the first-level input metrics include at least one second-level input metric, and the second-level input metrics may include at least one third-level input metric.

[0073] Exemplarily, the output metrics may be divided into two-level output metrics; for example, the output metrics may include first-level output metrics and second-level output metrics; the first-level output metrics may include at least one second-level output metric. For example, the first-level output metric is by product mileage, and the second-level output metrics may include mileage in the first stage (e.g., design stage), mileage in the second stage (e.g., development stage), mileage in the third stage (e.g., testing stage), and mileage in the fourth stage (e.g., sales stage). Of course, the output metrics may also be divided into three levels or more levels of output metrics; for example, the output metrics may include first-level output metrics, second-level output metrics, and third-level output metrics, where the first-level output metrics include at least one second-level output metric, and the second-level output metrics may include at least one third-level output metric.

[0074] Exemplarily, the first input metric and the first output metric may be a first-level input metric and a first-level output metric respectively; or, the first input metric and the first output metric may be multi-level input metrics and multi-level output metrics respectively.

[0075] Optionally, both the first data and the second data may be data corresponding to metrics (such as input metrics and / or output metrics).

[0076] Exemplarily, the first input index data in the first data is the data of the first input index; for example, if the first input index is R & D man-hours, then the first input index data is the time of R & D man-hours, such as 30 days; or if the first input index is funds, then the first input index data is the amount of funds, for example, 10,000 yuan.

[0077] Exemplarily, the first output index data in the first data is the data of the first output index; for example, if the first output index is the number of products, then the first output index data can be the quantity of the number of products, for example, 10,000 pieces; or if the first output index is product revenue, then the first output index data can be the amount of product revenue, for example, 1 million yuan.

[0078] Exemplarily, the second input index data in the second data is the data of the second input index; the second output index data in the second data is the data of the second output index.

[0079] Optionally, the first index may include at least the second index; the second index may include at least the third index. Exemplarily, the first input index may include at least the second input index; the first output index may include at least the second output index.

[0080] Optionally, the performance evaluation model may include any performance evaluation model; there is no limitation on this performance evaluation model, as long as this performance evaluation model can calculate the performance (or efficiency) of input and output. Exemplarily, the performance evaluation model can be, but is not limited to, a model based on Slack - Based Measure Data Envelopment Analysis (SBM - DEA), a super - efficiency DEA model, a non - radial DEA model, or a super - efficiency Slack - based Measure (SBM) model, etc.

[0081] Optionally, the linear regression model can be replaced by a panel regression model; there is no limitation on this linear regression model, as long as this linear regression model can determine the quantitative relationship of interdependence between two or more variables. This linear regression model can refer to a linear statistical analysis method. For example, this linear regression model can be a mixed - effects model, a fixed - effects model, or a random - effects model, etc.; or this linear regression model can be a multiple linear regression model, a logistic regression model, or a simple linear regression, etc.

[0082] Optionally, the significance test refers to the test of significance; the significance parameter is used to indicate whether it is significant or not; the significance parameter is used to indicate the magnitude of the difference between the regression coefficient and zero.

[0083] In some embodiments, step S101 may include: determining at least one product department with homogeneity to be evaluated. Here, the product departments with homogeneity may refer to those with relatively equivalent conditions in terms of product decision-making, product generation, and / or product sales.

[0084] In other embodiments, step S101 may include: selecting any one or more departments as at least one product department to be evaluated.

[0085] In still other embodiments, step S101 may include: selecting at least one core department as at least one product department to be evaluated. For example, the R & D department, sales department, and testing department, etc. are core departments.

[0086] In some embodiments, step 102 may include: selecting a second input indicator from the first input indicators based on the first input indicator data of the first input indicators of at least one product department; selecting a second output indicator from the first output indicators based on the first output indicator data of the first output indicators of at least one product department.

[0087] In some embodiments, inputting the second data of the second indicators of at least one product department into the efficiency evaluation model may include: inputting the second input indicator data of the second input indicators and the second output indicator data of the second output indicators of at least one product department into the efficiency evaluation model.

[0088] In the embodiments of the present invention, a linear regression model can be constructed by taking the efficiency result as the dependent variable and the second input indicator as the independent variable, so as to effectively analyze the influencing factors of the R & D efficiency of at least one department's products based on the linear regression model, and thus obtain the third input indicator with significant efficiency influence. In this way, on the one hand, the evaluation of efficiency in the R & D process is considered, which is convenient for intuitively evaluating the high and low R & D efficiency of multiple product departments within the enterprise. On the other hand, the input indicators can be calculated through the linear regression model, reducing the influence of subjective factors in efficiency evaluation and improving the accuracy of efficiency evaluation; moreover, statistical analysis can also be carried out on the influencing factors of R & D efficiency, which is convenient for providing effective suggestions for subsequent improvement and optimization of R & D efficiency.

[0089] In some embodiments, before step S102, it may include: obtaining first primary selection indicators of at least one product department to be evaluated, where the first primary selection indicators include first primary selection input indicators and first primary selection output indicators; the first primary selection indicators may be indicators selected from all the indicators of at least one product department according to historical experience or user selection, etc. Here, the first primary selection indicators are representative indicators among all the indicators of at least one product department, and the representative indicators may be indicators with corresponding data (such as specific time of R & D man-hours), indicators selected from historical experience, etc. Optionally, the first initial indicators may be used as the first indicators.

[0090] In some embodiments, before step S102, it may include: preprocessing the first initial indicators to obtain second initial indicators; the second primary selection indicators include second primary selection input indicators and second primary selection output indicators. Here, the preprocessing may be missing data processing and index value conversion, etc. Missing data processing may be filling in the missing data corresponding to the indicators; for example, a certain indicator has corresponding data in the 1st to 10th months and the 12th month, but no data in the 11th month; then the data for the 11th month can be filled in, such as the data for the 11th month can be the average or median value of the data of other months. Index value conversion may be converting the data corresponding to a certain indicator into data with unified dimensions; for example, the data of a certain indicator in the 1st to 9th months, the 11th month, and the 12th month are all * pieces (such as 1000 pieces, 1500 pieces), while in the 10th month it is * ten thousand pieces (such as 0.1 ten thousand pieces), then the * ten thousand pieces (such as 0.1 ten thousand pieces) in the 10th month can be converted into * pieces (such as 1000 pieces). Optionally, the second initial indicators may be used as the first indicators.

[0091] As Figure 2 shown, in some embodiments, step 102 includes:

[0092] Step S1021: Conduct a distinguishability test on the first data to obtain first indicators with a distinguishability greater than or equal to a predetermined distinguishability as first candidate indicators; among them, the first candidate indicators include first candidate input indicators and first candidate output indicators;

[0093] Step S1022: Conduct a redundancy test on the first candidate indicators to obtain first candidate indicators with a correlation coefficient less than or equal to a predetermined correlation coefficient as the second indicators; where the correlation coefficient is used to characterize the correlation between the first candidate indicators.

[0094] In some embodiments, step S1021 includes:

[0095] Determine the standard deviation of the first data of the first indicators for each product department; wherein, the number of the first indicators is the number of the standard deviations, the average value of the first data of each first indicator is the average value of the standard deviations, and the first data of each first indicator is the data point of the standard deviation;

[0096] Determine that the standard deviation is the recognition degree for performing a recognition test on the first data;

[0097] If the recognition degree of the product department is greater than or equal to the predetermined recognition degree, determine the first indicator of the product department as the first candidate indicator.

[0098] Optionally, the recognition degree can be the standard deviation of the first data.

[0099] Optionally, the recognition degree is used to represent the difference of each first indicator; the recognition degree is positively correlated with the magnitude of the difference. For example, the greater the recognition degree, the greater the difference; the smaller the recognition degree, the smaller the difference.

[0100] Optionally, the predetermined recognition degree can be any value, or can be a value set according to historical experience, etc. For example, if the recognition degree range is from 0 to 1, the predetermined recognition degree can be a value greater than or equal to 0.6.

[0101] Optionally, determining the standard deviation of the first data of the first indicators for each product department may include: determining a first value based on the average value of the first data of each of the first indicators; determining a second value corresponding to each of the first values based on the difference between each of the first data and the first value; determining a third value based on the sum of each of the second values; determining a fourth value based on the ratio of the third value to the number of the first indicators; determining the standard deviation of the first data based on the square root of the fourth value.

[0102] Exemplarily, the standard deviation of the first data of the first indicators for each product department can be: wherein, n is the number of the first indicators; x i is the first data of the i-th first indicator; is the average value of the first data of n first indicators; is the summation formula, which means the sum from the 1st to the nth.

[0103] In this way, in the embodiments of the present invention, indicators with relatively small standard deviations (i.e., recognition degrees) can be removed, because no matter how much is invested in these indicators, they are less likely to reflect the change of the investment; thus, after removing these indicators, the accuracy of determining the efficiency result subsequently can be improved.

[0104] In some embodiments, the redundancy test of the first candidate metrics in step S1022 means: performing a redundancy test on the first candidate input metric data of the first candidate input metric and the first candidate output metric data of the first candidate output metric.

[0105] Optionally, a redundancy test and a correlation test. Here, the correlation coefficient can characterize the correlation or redundancy between the first candidate metrics.

[0106] Optionally, the correlation coefficient is positively correlated with the correlation. Exemplarily, the relatively larger the correlation coefficient, the stronger the correlation; the relatively smaller the correlation coefficient, the relatively weaker the correlation.

[0107] Optionally, the predetermined correlation coefficient can be any value, or can be a value set according to historical experience, etc. For example, if the correlation coefficient ranges from 0 to 1, the predetermined correlation coefficient can be a value greater than or equal to 0.6.

[0108] Thus, in the embodiments of the present invention, since metrics with relatively strong correlations have similar effects on performance, for metrics with relatively strong correlations, one of them can be removed or selected as the metric input to the performance evaluation model, which can also improve the accuracy of determining the performance result.

[0109] In the embodiments of the present invention, by performing a discriminability test and a redundancy test on the metrics, the data of the metrics input to the performance evaluation model can be determined, without inputting the data of all metrics into the performance evaluation model, thereby reducing the computational complexity and saving computational resources, etc.

[0110] In some embodiments, the performance evaluation model includes a first performance evaluation model and a second performance evaluation model. The first performance evaluation model is a non-linear model, and the second performance evaluation model is a non-linear evaluation model;

[0111] As Figure 3 shown, step S103 includes:

[0112] Step S1031: Input the quantity of the second input metrics, the second input metric data of each second input metric, the quantity of the second output metrics, and the second output metric data of each second output metric into the first performance evaluation model;

[0113] Step S1032: Convert the first performance evaluation model into the second performance evaluation model;

[0114] Step S1033: Based on the second performance evaluation model, obtain the performance results of at least one product department.

[0115] Optionally, step S1031 may include: determining the j-th fifth value based on the ratio of the j-th first slack variable to the second input index data of the j-th second input index of the k-th decision-making unit; determining the sixth value based on the mean of the fifth values from the 1st to the m-th; determining the seventh value based on the difference between the first constant and the first value; determining the r-th eighth value based on the ratio of the r-th second slack variable to the second output index data of the k-th second output index; determining the ninth value based on the mean of the eighth values from the 1st to the q-th; determining the tenth value based on the sum of the first constant and the ninth value; determining the minimum value of the efficiency result based on the ratio of the seventh value to the tenth value; wherein, the first slack variable refers to the slack variable of the input index (such as the slack variable of the second input index), and the second slack variable refers to the slack variable of the output index (such as the slack variable of the second output index); m is the number of second input indexes, j is greater than 0 and less than or equal to m; q is the number of second output indexes, r is greater than 0 and less than or equal to q; k is the decision-making unit, k = 1, 2,..., n; one unit represents a product department. Exemplarily, determine the first efficiency evaluation model: where ρ is the efficiency result; min is the formula for obtaining the minimum value; ∑ is the summation formula; X 、 x represents the input index; Y 、 y represents the output index; λ is the eigenvalue.

[0116] Optionally, step S1032 may include: determining the second efficiency result based on the minimum value of the difference between the third slack variable and the sixth value. Here, the formula determined in step S1032 may refer to the second efficiency evaluation model. Exemplarily, determine the second efficiency evaluation model: The model includes the transpose of the input matrix X n×m , the transpose of the expected output , the transpose of the non-expected output ; the model parameters mainly include the projection variable Λ, the slack variables S - , S g , S b , t; wherein, t may be the third slack variable; S g is the expected output; S b is the non-expected output; s 1 is the number of indicators of the expected output (such as the second output index), s 2 is the number of indicators of the non-expected output (such as the second output index).

[0117] In the embodiments of the present invention, by constructing the input indexes and output indexes of the product department to construct the efficiency model, both the evaluation of the R & D process efficiency and the evaluation of the R & D result benefit are considered; thus, an accurate efficiency result can be determined.

[0118] In some embodiments, the method includes:

[0119] Determine a window period for the second indicator of at least one product department;

[0120] Determine the second indicator within a window period as the second candidate indicator, where the second candidate indicator includes a second candidate input indicator corresponding to the second input indicator and a second candidate output indicator corresponding to the second output indicator;

[0121] Step S1031 includes: inputting the quantity of the second candidate input indicators, the second candidate input indicator data of each second candidate input indicator, the quantity of the second candidate output indicators, and the second candidate output indicator data of each second candidate output indicator into the first efficiency evaluation model.

[0122] Optionally, determine a predetermined time period as a window period; this window period can be based on hours, days, months, or quarters, etc.

[0123] Exemplarily, the second indicator can be a monthly indicator for each month, and the second indicator is the monthly indicator from the 1st to the 12th month; a window period can include multiple months, for example, 3 months; then the second candidate indicators can be Window 1 (including January, February, and March), Window 2 (February, March, and April), Window 3 (March, April, and May), …… Window 9 (September, October, and November), and Window 10 (October, November, and December); thus, each time the window slides, the decision-making unit at the earliest time node (for example, this time node can be a month) is removed from the window, and after the last time node in the previous window period, a new time node is added in chronological order.

[0124] In the embodiments of the present invention, the efficiency (or effectiveness) of each product department at each time node within the window can be calculated, so as to be able to consider the problem of the lag of the output of each product department to a certain extent.

[0125] In some embodiments, before constructing the linear regression model, it includes: performing at least one of the following tests on the second input indicator data of the second input indicator to obtain the second input indicator that passes the test: stationarity test, Granger causality test, and multicollinearity test;

[0126] Step S104 includes: using the efficiency result as the dependent variable and the second input indicator data of the second input indicator that passes the test as the independent variable to construct a linear regression model.

[0127] Optionally, the second input index that passes the test refers to the second input index that passes the stationarity test, the Granger causality test, and / or the multicollinearity test. Exemplarily, the stationarity test refers to testing whether the independent variable is stationary. Exemplarily, the Granger causality test mainly tests whether there is a causal relationship between the independent variable and the dependent variable. Exemplarily, the multicollinearity test is used to test whether the variables (independent variables and / or dependent variables) are multicollinear variables, where the multicollinear variables are variables that are collinear with at least one other variable.

[0128] In an embodiment of the present invention, the independent variable that passes the test can be used as an influencing factor index of the linear regression model, so that the linear regression model has higher stability.

[0129] In some embodiments, the method further includes:

[0130] Analyzing at least one of the regression coefficient, positive and negative nature, P-value, and significance of the third input index of at least one product department to obtain the influence result of each third input index on at least one product department, where the influence result characterizes the magnitude of the influence degree and / or the magnitude of the influence correlation;

[0131] The regression coefficient is a parameter indicating the magnitude of the influence of the independent variable on the dependent variable;

[0132] The positive and negative nature includes positive correlation or negative correlation. Positive correlation is a parameter indicating the magnitude of the increase of the dependent variable as the independent variable increases, and negative correlation is a parameter indicating the decrease of the dependent variable as the independent variable increases;

[0133] The P-value is used to indicate the significance of the regression coefficient;

[0134] Significance is used to indicate the difference between the regression coefficient and zero; significance is represented by the P-value.

[0135] Optionally, the regression coefficient may include a positive regression coefficient and / or a negative regression coefficient; the positive regression coefficient is used to indicate positive correlation; the negative regression coefficient is used to indicate negative correlation.

[0136] In an embodiment of the present invention, by analyzing at least one of the regression coefficient, positive and negative nature, P-value, and significance of each influencing factor (such as the third input index), the influence degree and relationship of the R & D efficiency of each product department can be deduced; thus, suggestions for improving and optimizing the efficiency corresponding to each product department can be provided.

[0137] To further explain any embodiment of the present invention, a specific embodiment is provided below:

[0138] As Figure 4 shown, the present invention proposes a method for processing R & D efficiency, which is executed by a terminal and may include the following steps:

[0139] Step S201: Screen the product department objects to be evaluated.

[0140] In this step S201, the departments that need to be evaluated for R & D efficiency are mainly screened. Considering the necessary comparability of each department, it is recommended to select product departments with homogeneity to form a set of objects to be evaluated. Here, homogeneity means that their product decision-making, product production, and / or product sales, etc. have relatively equal conditions. Here, the set of objects to be evaluated can be at least one product department in the previous embodiments.

[0141] Step S202: Construct indicators and collect sample data.

[0142] This step S202 includes three sub - processes: preliminary selection of the indicator system, collection and pre - processing of indicator data, and inspection and improvement of the indicator system.

[0143] When initially selecting the input and output indicator systems for product R & D, an organic combination of the comprehensive method and the analysis method is adopted for the initial selection of the evaluation indicator system. First, select some representative existing indicator systems related to product R & D input and output. Through analysis and comparison, summarize the alternative evaluation indicator systems. Second, according to the evaluation indicator requirements of the data envelopment analysis method, use the analysis method to gradually subdivide the evaluation purpose level by level until the second - level indicator layer. Third, combined with the alternative evaluation indicator systems in the first step, select or modify and adjust appropriate indicators as the specific indicator layer of the third level to form the final initially selected indicators. Here, the initially selected indicators can be the first initial indicators in the previous embodiments; the sample data can be the data of the indicators in the previous embodiments.

[0144] When collecting indicator sample data, extract the original data from the original databases of information systems such as the product R & D system and the sales system. After collecting the original data, further pre - processing work needs to be carried out on it, mainly including two parts: missing data processing and indicator value conversion and calculation. In the part of missing value processing, generally, effective estimation is carried out through other relevant indicators; in the part of indicator value conversion and calculation, methods such as data normalization are needed to keep the dimension of the indicators unified. When inspecting and improving the indicator system, it is necessary to further use quantitative analysis methods to test the necessity of the initially selected indicators. Common indicator system necessity test methods include indicator discrimination test and redundancy test. Among them, discrimination refers to the ability and effect of a statistical evaluation indicator in distinguishing the value characteristics of each evaluation unit in a certain aspect. Since the indicator samples have been dimensionless through data normalization, the standard deviation can be used for the discrimination test. The specific formula for the standard deviation is as follows: n is the number of indicators; x i is the data of the i - th indicator; is the average value of the data of n indicators; is the summation formula, which refers to the sum from the 1st to the nth.

[0145] Redundancy refers to the degree of repetition in the calculation content among the sub - evaluation indicators within the evaluation index system. The commonly used test method is the maximum irrelevance method. This method first calculates the correlation coefficients of each indicator variable to form a correlation coefficient matrix, calculates the correlation coefficient of each variable with the remaining variables, and then eliminates variables with too high correlation by setting a reasonable correlation coefficient threshold. Here, the indicators after the discrimination test and redundancy test can be the first indicator or the second initial indicator in the previous embodiments.

[0146] Step S203, establish an effectiveness evaluation model.

[0147] Based on the R & D investment, product indicators and sample data in the above - mentioned links, evaluate the product R & D effectiveness of each department by constructing a window data envelopment analysis model.

[0148] Data envelopment analysis method is a representative method in non - parametric analysis methods of statistical comprehensive evaluation, and is widely applicable to the effectiveness evaluation among departments with homogeneity. It is based on relative efficiency and the Pareto optimization theory in economics, and uses the linear programming theory in mathematics as the main technical means. For several decision - making units with multiple inputs and multiple outputs and with homogeneity, calculate the relative efficiency of each decision - making unit to conduct relative effectiveness evaluation, which is a relatively common system analysis and comprehensive evaluation method.

[0149] This solution uses the "super - efficiency" model (super efficiency model, SEA), that is, the SBM - DEA method for R & D effectiveness evaluation. Compared with the general data envelopment analysis model, the advantage of the SBM model is that it solves the problem that the radial model does not include slack variables in the measurement of inefficiency. The specific model is as follows: Among them, the objective function ρ represents the R & D effectiveness value (that is, the effectiveness result in the previous embodiments); for each decision - making unit k = 1, 2,, n. At this time, this model is a non - linear model. Convert this model into a linear model and add non - desired outputs to the model to get: The model parameters mainly include the projection variable Λ, slack variables S - , S g , S b , t.

[0150] The window data envelopment analysis model is a commonly used panel data analysis method. All product departments within a certain time period (window period) of a certain width are selected as the reference set for one evaluation. When actually calculating using the window SBM-DEA method, each time the window slides, the decision-making unit at the earliest time node is removed from the window, and after the last time node in the previous window period, a new time node is added in chronological order.

[0151] Based on the above window data envelopment analysis model, calculate the efficiency of each product department at each time node within the window in turn, and by calculating the average efficiency of each product department in each window period, take it as the actual efficiency of the current period. This method also considers to a certain extent the lag problem of the output of each product department.

[0152] Step S204, output the performance evaluation result.

[0153] Combined with the historical product R & D investment and output index data collected from each product department, perform the calculation of the window data envelopment analysis model, and finally obtain the R & D performance size results of each product department; this index is a single index, that is, the performance level index of the department.

[0154] Step S205, construct performance influencing factor indicators and collect sample data.

[0155] In the product R & D investment index set obtained in the above step S202, combined with the existing research theories on R & D performance influencing factors, select some representative indicators as performance influencing indicators, and collect the corresponding sample data. Here, the performance influencing indicators can be the second input indicators in the previous embodiments.

[0156] Step S206, establish a performance influencing factor analysis model.

[0157] Based on the R & D performance evaluation result data of each department in different periods obtained in the above step S204, set it as the dependent variable. At the same time, combined with the influencing factor indicators and sample data obtained in the above step S205, set them as the independent variables. Through constructing a suitable panel data regression model, effectively analyze the influencing factors of the R & D performance of each department. The specific process is as follows:

[0158] (1) Data inspection.

[0159] When selecting a specific panel data regression model, it is necessary to conduct necessary basic inspections on the existing sample data and variables, usually including three parts: the stationarity test of variables, the Granger causality test between explanatory variables and explained variables, and the multicollinearity among explanatory variables.

[0160] First, perform a stationarity test. Some non-stationary economic time series often show common trends of change, and there may not necessarily be a direct correlation between these series themselves. To avoid the phenomenon of spurious regression, before establishing a model, it is necessary to perform a stationarity test on the variable data that may be used in the model. Generally speaking, before performing a stationarity test on a series, the series can be logarithmized, which can bring benefits such as reducing collinearity and heteroscedasticity.

[0161] Second, perform a Granger causality test. After determining the stationarity of each variable in the model, a Granger causality test can be further performed on the panel variable data to explore whether there is a causal relationship between the explanatory variable and the explained variable.

[0162] Third, perform a multicollinearity test. Before selecting a model, a simple multiple linear regression can be established to calculate the variance inflation factor VIF to test the multicollinearity between the explanatory variables.

[0163] Finally, non-stationarity indicators and influencing factor indicators that fail the Granger causality test and multicollinearity test are proposed as the final influencing factor indicators for the model.

[0164] (2) Select an appropriate panel data regression model.

[0165] Panel regression models usually have three forms. One is the pooled estimation model, the second is the fixed effects model, and the third is the random effects model. If the intercept term in the fixed effects model includes the average effects of the cross-sectional random error term and the time random error term, and both of these random error terms follow a normal distribution, then the fixed effects model becomes a random effects model. Generally speaking, the F test, Hausman test, and LM test can be used to further determine which type of panel data model is suitable. It is only necessary to select a specific form of the panel regression data model in combination with the test results, and finally calculate and establish the regression model of the influencing factors of the efficiency of each product department, and estimate the parameters of the regression coefficients of each influencing factor.

[0166] Step S207, output suggestions for efficiency improvement and optimization.

[0167] Combined with the results of the regression model of the efficiency influencing factors established in step S206 above, by analyzing the magnitudes, positive and negative signs, P values, significance levels, etc. of the regression coefficients of each influencing factor, it is possible to further deduce the influence degrees and relationships of each influencing factor on the R & D efficiency of the product department, and then put forward corresponding suggestions for efficiency improvement and optimization for each product department.

[0168] In an embodiment of the invention, a method for evaluating R & D efficiency and analyzing influencing factors based on a fusion model is proposed. In the R & D efficiency evaluation stage, by constructing input indicators and output indicators for product R & D, both the evaluation of R & D process efficiency and the evaluation of R & D result benefits are considered. At the same time, a window data envelopment analysis evaluation model is constructed to finally evaluate the efficiency of product departments. Compared with traditional efficiency evaluation methods, it effectively solves the problem of subjective factor influence in the construction of efficiency evaluation indicators, and can more intuitively display the efficiency levels of each department through individual efficiency evaluation results.

[0169] Moreover, in the stage of analyzing influencing factors of efficiency, by constructing an influencing factor verification analysis, homogeneous or non-causal influencing factors are eliminated, effectively improving the accuracy of influencing factor analysis. At the same time, the result of the efficiency evaluation model is fused with the final influencing factor regression analysis model; therefore, when designing the dependent variable, instead of comprehensively calculating multiple efficiency indicators according to different weights, the evaluation result of the window data envelopment analysis model is used as the dependent variable, and the established influencing factor regression model has higher robustness and accuracy.

[0170] It should be noted here that the following description of the R & D efficiency processing device is similar to the description of the above R & D efficiency processing method, and the beneficial effects of the method will not be elaborated. For the technical details not disclosed in the embodiment of the R & D efficiency processing device of the present invention, please refer to the description of the embodiment of the R & D efficiency processing method of the present invention.

[0171] As Figure 5 shown, an embodiment of the present invention provides an R & D efficiency processing device, including:

[0172] A determination module 31, configured to determine at least one product department to be evaluated;

[0173] A first processing module 32, configured to select a second indicator from the first indicators based on the first data of the first indicators of at least one product department; wherein, the first indicators include first input indicators and first output indicators; the first data includes first input indicator data of the first input indicators and first output indicator data of the first output indicators; the second indicators include second input indicators selected from the first input indicators and second output indicators selected from the first output indicators;

[0174] A second processing module 33, configured to input the second data of the second indicators of at least one product department into an efficiency evaluation model to obtain an efficiency result of at least one product department; wherein, the second data includes second input indicator data of the second input indicators and second output indicator data of the second output indicators; the efficiency result is used to indicate the efficiency ranking of at least one product department;

[0175] The second processing module 33 is configured to use the efficiency result as the dependent variable and the second input index data of the second input index as the independent variable to construct a linear regression model;

[0176] The second processing module 33 is configured to perform a significance test on the independent variable based on the linear regression model to obtain the independent variable with a significance parameter higher than the threshold of the significance test as the third input index.

[0177] In some embodiments, the first processing module 32 is configured to:

[0178] Perform a recognition test on the first data to obtain the first index with a recognition greater than or equal to a predetermined recognition as the first candidate index; wherein, the first candidate index includes a first candidate input index and a first candidate output index;

[0179] Perform a redundancy test on the first candidate index to obtain the first candidate index with a correlation coefficient less than or equal to a predetermined correlation coefficient as the second index; wherein, the correlation coefficient is used to characterize the correlation between the first candidate indexes.

[0180] In some embodiments, the first processing module 32 is configured to:

[0181] Determine the standard deviation of the first data of the first index of each product department; wherein, the number of the first indexes is the number of the standard deviations, the average value of the first data of each first index is the average value of the standard deviations, and the first data of each first index is the data point of the standard deviation;

[0182] Determine that the standard deviation is the recognition for performing the recognition test on the first data;

[0183] If the recognition of the product department is greater than or equal to the predetermined recognition, determine the first index of the product department as the first candidate index.

[0184] In some embodiments, the efficiency evaluation model includes a first efficiency evaluation model and a second efficiency evaluation model, the first efficiency evaluation model is a non-linear model, and the second efficiency evaluation model is a non-linear evaluation model; the second processing module 33 is configured to:

[0185] Input the number of the second input indexes, the second input index data of each second input index, the number of the second output indexes, and the second output index data of each second output index into the first efficiency evaluation model;

[0186] Convert the first efficiency evaluation model into the second efficiency evaluation model;

[0187] Based on the second efficiency evaluation model, obtain the efficiency results of at least one product department.

[0188] In some embodiments, a determination module 31 is configured to determine a window period for a second indicator of at least one product department; determine the second indicator within a window period as a second candidate indicator, where the second candidate indicator includes a second candidate input indicator corresponding to a second input indicator and a second candidate output indicator corresponding to a second output indicator;

[0189] A second processing module 33 is configured to input the quantity of the second candidate input indicators, the second candidate input indicator data of each second candidate input indicator, the quantity of the second candidate output indicators, and the second candidate output indicator data of each second candidate output indicator into a first efficiency evaluation model.

[0190] In some embodiments, the second processing module 33 is configured to perform at least one of the following tests on the second input indicator data of the second input indicator to obtain a second input indicator that passes the test: stationarity test, Granger causality test, and multicollinearity test;

[0191] The second processing module 33 is further configured to use the efficiency result as a dependent variable and the second input indicator data of the second input indicator that passes the test as independent variables to construct a linear regression model.

[0192] In some embodiments, a first processing module 32 is configured to perform at least one of the analysis of the regression coefficient, positive or negative nature, P-value, and significance of the third input indicator of at least one product department to obtain the influence result of each third input indicator on at least one product department, where the influence result characterizes the magnitude of the influence degree and / or the magnitude of the influence correlation;

[0193] The regression coefficient is a parameter indicating the magnitude of the influence of an independent variable on a dependent variable;

[0194] The positive or negative nature includes positive correlation or negative correlation. Positive correlation is a parameter indicating the magnitude of the increase of the dependent variable as the independent variable increases, and negative correlation is a parameter indicating the decrease of the dependent variable as the independent variable increases;

[0195] The P-value is used to indicate the significance of the regression coefficient;

[0196] Significance is used to indicate the difference between the regression coefficient and zero; significance is represented by the P-value.

[0197] As Figure 6 shown, an embodiment of the present invention further provides a terminal, where the terminal includes a processor 41 and a memory 42 for storing a computer program that can run on the processor 41; wherein, when the processor 41 is configured to run the computer program, the R & D efficiency processing method according to any embodiment of the present invention is implemented.

[0198] In some embodiments, the memory in the embodiments of the present invention may be a volatile memory or a non-volatile memory, or may include both volatile and non-volatile memories. Among them, the non-volatile memory may be a read-only memory (ROM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), an electrically erasable programmable read-only memory (EEPROM), or a flash memory. The volatile memory may be a random access memory (RAM), which is used as an external cache. By way of example but not limitation, many forms of RAM are available, such as static random access memory (SRAM), dynamic random access memory (DRAM), synchronous dynamic random access memory (SDRAM), double data rate synchronous dynamic random access memory (DDR SDRAM), enhanced synchronous dynamic random access memory (ESDRAM), synchlink dynamic random access memory (SLDRAM), and direct rambus random access memory (DRRAM). The memory of the systems and methods described herein is intended to include but not be limited to these and any other suitable types of memory.

[0199] The processor may be an integrated circuit chip with the ability to process signals. In the implementation process, each step of the above method can be completed by the integrated logic circuit in the hardware of the processor or by instructions in the form of software. The above-mentioned processor may be a general-purpose processor, a digital signal processor (DSP), an application specific integrated circuit (ASIC), a field programmable gate array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components. It can implement or execute the various methods, steps, and logic block diagrams disclosed in the embodiments of the present invention. The general-purpose processor may be a microprocessor, or the processor may also be any conventional processor, etc. The steps of the method disclosed in combination with the embodiments of the present invention can be directly embodied as being executed and completed by a hardware decoding processor, or by a combination of hardware and software modules in the decoding processor. The software module may be located in a mature storage medium in the art such as a random access memory, a flash memory, a read-only memory, a programmable read-only memory, or an electrically erasable programmable memory, a register, etc. This storage medium is located in the memory, and the processor reads the information in the memory and combines its hardware to complete the steps of the above method.

[0200] In some embodiments, the embodiments described herein can be implemented using hardware, software, firmware, middleware, microcode, or a combination thereof. For a hardware implementation, the processing unit can be implemented in one or more application specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field-programmable gate arrays (FPGAs), general-purpose processors, controllers, microcontrollers, microprocessors, other electronic units for performing the functions described in the invention, or a combination thereof.

[0201] For a software implementation, the techniques described herein can be implemented by modules (e.g., procedures, functions, etc.) that execute the functions described herein. The software code can be stored in a memory and executed by a processor. The memory can be implemented inside or outside the processor.

[0202] An embodiment of the present invention provides a computer storage medium. The computer-readable storage medium stores an executable program. When the executable program is executed by a processor, the steps of the R & D efficiency processing method according to any embodiment of the present invention can be implemented.

[0203] An embodiment of the present invention provides a computer program product. The computer program product includes a computer program or instruction. When the computer program or instruction is executed by a processor, the steps of the R & D efficiency processing method according to any embodiment of the present invention are implemented.

[0204] In some embodiments, the computer storage medium may include: various media capable of storing program codes, such as USB flash drives, mobile hard disks, read-only memories (ROMs), random access memories (RAMs), magnetic disks, or optical discs.

[0205] It should be noted that: between the technical solutions described in the embodiments of the present invention, any combination can be made without conflict.

[0206] As described above, the above are only specific embodiments of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art within the technical scope disclosed by the present invention can easily think of changes or substitutions, which should all be covered within the protection scope of the present invention. Therefore, the protection scope of the present invention should be subject to the protection scope of the claims.

Claims

1. A method for processing R&D efficiency, characterized in that: include: Identify at least one product segment to be evaluated; Based on the first data of the first indicator of the at least one product department, a second indicator is selected from the first indicator; wherein the first indicator includes a first input indicator and a first output indicator; the first data includes first input indicator data of the first input indicator and first output indicator data of the first output indicator; the second indicator includes a second input indicator selected from the first input indicator and a second output indicator selected from the first output indicator; Inputting the second data of the second indicator of the at least one product department into the performance evaluation model to obtain the performance result of the at least one product department; wherein the second data includes the second input indicator data of the second input indicator and the second output indicator data of the second output indicator; and the performance result is used to indicate the performance ranking of the at least one product department; Using the performance result as a dependent variable and the second input indicator data of the second input indicator as an independent variable to construct a linear regression model; Based on the linear regression model, a significance test is performed on the independent variable to obtain the independent variable whose significance parameter of the significance test is higher than a threshold value as the third input indicator.

2. The method according to claim 1, characterized in that The selecting a second indicator from the first indicator based on the first data of the first indicator of the at least one product department comprises: Performing a recognition test on the first data to obtain the first indicator having a recognition greater than or equal to a predetermined recognition as a first candidate indicator; wherein the first candidate indicator includes a first candidate input indicator and a first candidate output indicator; A redundancy check is performed on the first candidate indicators to obtain the first candidate indicators whose correlation coefficients are less than or equal to a predetermined correlation coefficient as the second indicators; wherein the correlation coefficients are used to characterize the correlation between the first candidate indicators.

3. The method according to claim 2, characterized in that The performing a recognition test on the first data to obtain the first indicator having a recognition greater than or equal to a predetermined recognition as the first candidate indicator includes: Determine the standard deviation of the first data of the first indicator of each product department; wherein the number of the first indicators is the number of the standard deviations, the average value of the first data of each first indicator is the average value of the standard deviations, and the first data of each first indicator is a data point of the standard deviations; Determining the standard deviation as the recognition degree for performing a recognition degree test on the first data; If the recognition degree of the product department is greater than or equal to a predetermined recognition degree, the first indicator of the product department is determined as a first candidate indicator.

4. The method according to any one of claims 1 to 3, characterized in that: The performance evaluation model includes a first performance evaluation model and a second performance evaluation model, wherein the first performance evaluation model is a nonlinear model and the second performance evaluation model is a nonlinear evaluation model; inputting the second data of the second indicator of the at least one product department into the performance evaluation model to obtain the performance result of the at least one product department includes: Inputting the number of the second input indicators, the second input indicator data of each of the second input indicators, the number of the second output indicators, and the second output indicator data of each of the second output indicators into the first performance evaluation model; converting the first performance evaluation model into the second performance evaluation model; The performance result of the at least one product department is obtained based on the second performance evaluation model.

5. The method according to claim 4, characterized in that The method comprises: Determining a window period for the second indicator of the at least one product department; Determine the second indicator within a window period as a second candidate indicator, wherein the second candidate indicator includes a second candidate input indicator corresponding to the second input indicator and a second candidate output indicator corresponding to the second output indicator; The step of inputting the number of the second input indicators, the second input indicator data of each second input indicator, the number of the second output indicators, and the second output indicator data of each second output indicator into the first performance evaluation model includes: inputting the number of the second candidate input indicators, the second candidate input indicator data of each second candidate input indicator, the number of the second candidate output indicators, and the second candidate output indicator data of each second candidate output indicator into the first performance evaluation model.

6. The method according to any one of claims 1 to 3, characterized in that: Before constructing the linear regression model, it includes: Performing at least one of the following tests on the second input indicator data of the second input indicator to obtain the second input indicator that passes the test: a stationarity test, a Granger causality test, and a multicollinearity test; The taking the performance result as a dependent variable and the second input indicator data of the second input indicator as an independent variable to construct a linear regression model includes: The linear regression model is constructed by taking the performance result as a dependent variable and the second input indicator data of the second input indicator that has passed the test as an independent variable.

7. The method according to any one of claims 1 to 3, characterized in that: The method further comprises: Perform at least one of regression coefficient, positivity, P value and significance analysis on the third input indicator of the at least one product department to obtain the impact result of each third input indicator on the at least one product department, wherein the impact result represents the magnitude of the impact degree and / or the magnitude of the impact correlation; The regression coefficient is a parameter indicating the influence of the independent variable on the dependent variable. The positive and negative correlation includes positive correlation or negative correlation, wherein the positive correlation is used to indicate that the size parameter of the dependent variable increases as the independent variable increases, and the negative correlation is used to indicate that the size parameter of the dependent variable decreases as the independent variable increases; The P value is used to indicate the significance of the regression coefficient; The significance is used to indicate the difference between the regression coefficient and zero; the significance is represented by the P value.

8. A research and development efficiency processing device, characterized in that: include: a determination module for determining at least one product sector to be evaluated; A first processing module, configured to select a second indicator from the first indicator based on the first data of the first indicator of the at least one product department; wherein the first indicator includes a first input indicator and a first output indicator; the first data includes first input indicator data of the first input indicator and first output indicator data of the first output indicator; the second indicator includes a second input indicator selected from the first input indicator and a second output indicator selected from the first output indicator; a second processing module, configured to input second data of the second indicator of the at least one product department into a performance evaluation model to obtain a performance result of the at least one product department; wherein the second data includes second input indicator data of the second input indicator and second output indicator data of the second output indicator; and the performance result is used to indicate a performance ranking of the at least one product department; The second processing module is used to construct a linear regression model by taking the performance result as a dependent variable and the second input indicator data of the second input indicator as an independent variable; The second processing module is used to perform a significance test on the independent variable based on the linear regression model to obtain the independent variable whose significance parameter of the significance test is higher than a threshold as the third input indicator.

9. A terminal, characterized in that: The terminal includes a processor and a memory for storing a computer program that can be run on the processor; wherein, when the processor is used to run the computer program, the R&D efficiency processing method described in any one of claims 1 to 7 is implemented.

10. A computer storage medium, characterized in that: The computer storage medium contains computer executable instructions, wherein the computer executable instructions are executed by a processor to implement the R&D efficiency processing method according to any one of claims 1 to 7.

11. A computer program product, comprising a computer program or instructions, characterized in that: When the computer program or instruction is executed by a processor, the R&D efficiency processing method described in any one of claims 1 to 7 is implemented.

Images