Innovation performance evaluation method and system based on stepped efficiency optimization

Abstract

The invention discloses an innovation performance evaluation method and system based on stepped efficiency optimization, and the method comprises the steps: S100: obtaining to-be-evaluated data, carrying out the processing of the data, removing negative sample data and zero-index-containing data, and obtaining a decision object; s200, designing a classification index system according to the decision object, performing score assignment on a subdivision index, and performing normalization processing on the assignment; and S300, establishing a mixing efficiency evaluation model, a pure efficiency evaluation model and a second-order efficiency evaluation model, and performing evaluation based on the mixing efficiency evaluation model, the pure efficiency evaluation model and the second-order efficiency evaluation model. According to the method, the input and output elements of the innovation system are analyzed and sorted, the performance evaluation index system is constructed in a classified manner, the innovation efficiency is calculated and sorted by adopting a stepped segmentation efficiency optimization method, and the innovation performance value is calculated according to the scoring conversion rule, so that a quantitative scientific judgment conclusion of the innovation performance level of the scientific research project is formed.

Description

Technical Field

[0001] The present invention belongs to the technical field of scientific research information management and evaluation, and more specifically, relates to an innovation performance evaluation method and system for step-by-step efficiency optimization. Background Art

[0002] At present, with the continuous development of social science and technology, the complexity of innovation activities is also showing an accelerating growth trend. The complexity of innovation activities is manifested in the following aspects: First, multidisciplinary integration. Modern scientific research often involves multidisciplinary knowledge and technology, and innovation requires the integration of multidisciplinary knowledge contributions; second, the research cycle is long, and many major scientific research projects require years or even decades of continuous and in-depth research; third, there are many uncertain factors, the difficulty of innovation is constantly increasing, the innovation risk is difficult to quantify and capture the law, the contingency of the output of innovation activities is increasing, and the output path of scientific research results is difficult to solidify. At the same time, the complexity of innovation activities will also aggravate the "black box" effect of scientific research itself, that is, it will continue to amplify the negative effects of the increasingly difficult evaluation of innovation performance due to factors such as strong professional and technical barriers, opaque information, and many unobservable factors (such as the creativity and inspiration of scientific researchers). The above reasons make it difficult to accurately observe and scientifically evaluate the level of innovation performance, and it is impossible to reward the good and punish the bad according to the real level of innovation performance, which in turn affects the effective implementation of the innovation incentive mechanism.

[0003] In traditional research methods, there are many technical tools for innovation performance evaluation. Commonly used evaluation methods include Delphi method, hierarchical analysis method, brainstorming method, and objective evaluation methods include TOPSIS analysis method, grey correlation analysis method, variable weight analysis method, etc. Although these methods can conduct comprehensive, systematic and scientific evaluation of the performance level of innovation activities from different perspectives, the above methods have received mixed reviews and have different advantages and disadvantages in use. The most critical point is that the above methods do not fully reflect the "input-output" law characteristics in the research cycle of scientific research projects in the evaluation process. They only focus on scientific research output but ignore scientific research input. They cannot provide timely improvement feedback on innovation activities and cannot effectively adjust and optimize the number and proportion of scientific research input factors to obtain the maximum scientific research output efficiency. Summary of the invention

[0004] In view of the above defects or improvement needs of the prior art, the present invention provides an innovation performance evaluation method for step-by-step efficiency optimization. By analyzing the input and output elements of the innovation system, the "input-output efficiency" is extracted as the common goal of innovation performance evaluation, and a performance evaluation index system is constructed by classification. A hybrid efficiency evaluation model, a pure efficiency evaluation model and a second-order efficiency evaluation model are established to calculate and optimize the innovation efficiency. The innovation performance value is calculated according to the scoring conversion rules to form a quantitative evaluation of the innovation performance level.

[0005] To achieve the above object, according to the first aspect of the present invention, an innovative performance evaluation method for step - by - step efficiency optimization is provided, including:

[0006] S100: Obtain the data to be evaluated, process the data, eliminate negative sample data and data with zero indicators, and obtain decision - making objects;

[0007] S200: Design a classification index system according to the decision - making objects, score and assign values to the sub - indicators, and perform normalization processing on the assigned values;

[0008] S300: Establish a hybrid efficiency evaluation model, a pure efficiency evaluation model, and a second - order efficiency evaluation model, where:

[0009] When examining the innovation efficiency of scientific research projects without eliminating the influence of scale efficiency, perform efficiency evaluation based on the hybrid efficiency evaluation model;

[0010] When examining the innovation efficiency of scientific research projects and needing to eliminate the influence of scale efficiency, perform efficiency evaluation based on the pure efficiency evaluation model;

[0011] When there is a situation where the innovation efficiency values of multiple decision - making objects are all 1 in the evaluation results of the hybrid efficiency evaluation model or the pure efficiency evaluation model, perform evaluation based on the second - order efficiency evaluation model.

[0012] Further, in step S300, the establishment of the hybrid efficiency evaluation model includes:

[0013] S301: Determine a group of a total of n decision - making objects, and the target decision - making object to be measured is DMO k , denoted as DMO j (j = 1, 2, … n) in sequence, and its output - input ratio is expressed as:

[0014]

[0015] S302: Accordingly, obtain the hybrid efficiency evaluation model as:

[0016]

[0017]

[0018] Among them, the input types of each decision - making object are m, the output types are q, and their respective numerical values are x i (i = 1, 2, … m) and y r (r = 1, 2, … q), the input weights and output weights are v i (i = 1, 2, … m) and u r (r = 1, 2, … q); k ≤ n.

[0019] Further, in step S300, the establishment of the hybrid efficiency evaluation model includes: transforming the hybrid efficiency evaluation model, and letting:

[0020]

[0021] μ = tu,

[0022] γ = tv,

[0023] transform the hybrid efficiency evaluation model into an equivalent linear programming model:

[0024]

[0025]

[0026] Further, in step S300, the establishment of the hybrid efficiency evaluation model includes: by performing a dual transformation on the linear programming model, the final expression of the hybrid efficiency evaluation model can be obtained as:

[0027] minθ

[0028]

[0029] where λ represents the linear combination coefficient of the decision-making object, and the optimal solution θ of the model * represents the efficiency value, and satisfies the condition θ * ∈(0, 1].

[0030] Further, in step S300, the establishment of the pure efficiency evaluation model includes:

[0031] S303: When the scale of scientific and technological innovation changes and the impact degree is large and cannot be ignored, it is necessary to supplement and correct the hybrid efficiency evaluation model, and supplement the constraint condition The pure efficiency measurement model is established as:

[0032] minθ

[0033]

[0034] where λ ≥ 0.

[0035] Further, in step S300, the establishment of the pure efficiency evaluation model includes: by performing a dual transformation on the pure efficiency measurement model, the final expression of the pure efficiency measurement model can be obtained as:

[0036]

[0037]

[0038] Furthermore, in step S300, the establishment of the second-order efficiency evaluation model includes: based on the basic assumption that the input-oriented scale efficiency remains unchanged, supplementing the constraint condition j≠k, and transforming the hybrid efficiency evaluation model into a second-order efficiency evaluation model:

[0039] minθ

[0040]

[0041] Furthermore, in step S300, the establishment of the second-order efficiency evaluation model includes: when the scale efficiency based on input orientation becomes the basic assumption, the additional constraint condition The pure efficiency evaluation model is transformed into a second-order efficiency evaluation model:

[0042] minθ

[0043]

[0044] Furthermore, in step S100, the decision objects satisfy: the number n of decision objects is greater than or equal to three times the number of input and output indicators, and the number n of decision objects is greater than or equal to the product of the number of input and output indicators, that is:

[0045] n≥max{3×(m+q),m×q}

[0046] Among them, m and q are the number of scientific research input indicators and the number of scientific research output indicators respectively.

[0047] According to a second aspect of the present invention, a step-by-step efficiency optimization innovation performance evaluation system is provided, which is used to implement the step-by-step efficiency optimization innovation performance evaluation method, comprising:

[0048] The data acquisition module is used to obtain the data to be evaluated, process the data, remove negative sample data and data containing zero indicators, and obtain the decision object;

[0049] A data processing module is used to design a classification index system according to the decision object, assign scores to the subdivided indicators, and normalize the assigned scores;

[0050] The module establishment module is used to establish a mixed efficiency evaluation model, a pure efficiency evaluation model and a second-order efficiency evaluation model, where:

[0051] When examining the innovation efficiency of scientific research projects and without eliminating the impact of scale efficiency, efficiency evaluation is performed based on the hybrid efficiency evaluation model;

[0052] When examining the innovation efficiency of scientific research projects and eliminating the impact of scale efficiency, efficiency evaluation is performed based on the pure efficiency evaluation model;

[0053] When there is a situation where the innovation efficiency values of multiple decision-making objects are all 1 in the evaluation results of the mixed efficiency evaluation model or the pure efficiency evaluation model, the evaluation is carried out based on the second-order efficiency evaluation model.

[0054] Generally speaking, compared with the prior art, the above technical solutions conceived by the present invention can achieve the following beneficial effects:

[0055] 1. The method of the present invention analyzes the input and output elements of the innovation system, extracts "input-output efficiency" as the common goal of innovation performance evaluation, classifies and constructs a performance evaluation index system, and establishes a mixed efficiency evaluation model, a pure efficiency evaluation model, and a second-order efficiency evaluation model for innovation efficiency calculation and optimization. According to the scoring conversion rule, the innovation performance value is calculated to form a quantitative evaluation of the innovation performance level.

[0056] 1. The method of the present invention is a non-parametric method, which does not require presupposing the specific form of the scientific research input-output function in advance, avoids the interference of subjective factors on the function form assumption, and preferably circumvents the problem that it is difficult to scientifically quantify the innovation performance brought by the innovation "black box" effect. At the same time, by obtaining the relative efficiency differences between different projects through the mathematical programming model, the secondary discrimination of the innovation efficiency can be effectively improved, and the ultimate goal of step-by-step efficiency optimization can be achieved.

[0057] 3. The method of the present invention is more in line with the actual situation of multi-input and multi-output of innovation activities. When calculating the innovation performance and comparing the pros and cons of scientific research project performances, the dimensional and magnitude constraint conditions of each index element of input and output are relatively loose, and it is not necessary to set weights between each element index. Therefore, this technical method has wider applicability at the application level.

[0058] 4. The method of the present invention can not only provide the evaluation ranking of scientific research performance, but also feedback on the innovation process through the performance results, and provide specific improvement plans for the input and output of the elements of innovation activities. This method has strong practical guiding significance for scientifically and efficiently arranging the input quantity and direction of various scientific research resources such as human, financial, material, and time. BRIEF DESCRIPTION OF THE DRAWINGS

[0059] Figure 1 It is the structural diagram of the innovation performance evaluation principle for step-by-step efficiency optimization according to the embodiment of the present invention;

[0060] Figure 2 It is the schematic flowchart of the innovation performance evaluation method for step-by-step efficiency optimization according to the embodiment of the present invention;

[0061] Figure 3 It is the composition architecture diagram of the innovation performance evaluation system for step-by-step efficiency optimization according to the embodiment of the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0062] In order to make the purpose, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with 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 intended to limit the present invention. In addition, the technical features involved in the various embodiments of the present invention described below can be combined with each other as long as they do not conflict with each other.

[0063] Example 1

[0064] like Figure 1 and Figure 2 As shown, an embodiment of the present invention provides an innovation performance evaluation method of step-by-step efficiency optimization. The operation process of the entire model includes two sequentially connected stages of data indicator preparation and step-by-step efficiency optimization algorithm technology, including: S100: obtaining the data to be evaluated, and processing the data, eliminating negative sample data and data containing zero indicators, and obtaining a decision object; S200: designing a classification indicator system according to the decision object, and scoring and assigning values ​​to the subdivided indicators, and normalizing the assignments; S300: establishing a hybrid efficiency evaluation model, a pure efficiency evaluation model, and a second-order efficiency evaluation model, wherein: when the innovation efficiency of a scientific research project is examined and there is no need to eliminate the impact of scale efficiency, efficiency evaluation is performed based on the hybrid efficiency evaluation model; when the innovation efficiency of a scientific research project is examined and the impact of scale efficiency needs to be eliminated, efficiency evaluation is performed based on the pure efficiency evaluation model; when the innovation efficiency of a plurality of decision objects in the evaluation results of the hybrid efficiency evaluation model or the pure efficiency evaluation model are all 1, evaluation is performed based on the second-order efficiency evaluation model. The method of the present invention analyzes the input and output elements of the innovation system, extracts "input-output efficiency" as the common goal of innovation performance evaluation, classifies and constructs a performance evaluation index system, and establishes a hybrid efficiency evaluation model, a pure efficiency evaluation model and a second-order efficiency evaluation model to calculate and optimize the innovation efficiency, calculates the innovation performance value according to the scoring conversion rules, and forms a quantitative evaluation of the innovation performance level.

[0065] The specific implementation is described as follows according to the decomposition steps at the operational level:

[0066] Step 1: Select the evaluation object. Define the range of alternative objects for the scientific research projects to be evaluated. Conduct a preliminary inspection and screening according to the integrity of the sample data of the scientific research projects, and refine the alternative scientific research projects that meet the evaluation conditions as the decision-making objects (Decision-Making Object, abbreviated as DMO). And the number of decision-making objects should not be too small to avoid lacking the necessary discrimination when evaluating the efficiency of DMO using the method of this patent. Generally, the rule is that the number n of DMO should satisfy two conditions simultaneously: (1) The number of DMO is greater than or equal to 3 times the sum of the input and output index numbers; (2) The number of DMO is greater than or equal to the product of the input and output index numbers. That is

[0067] n≥max{3×(m+q),m×q} (1)

[0068] In formula (1), m and q are the numbers of scientific research input indicators and scientific research output indicators respectively.

[0069] Step 2: Prepare for data utilization. Conduct necessary processing on the sample data to improve the accuracy of data analysis for DMO. For the bad output scientific research input data, it can be processed in the form of negative numbers. For the situation where there is 0 in the indicators, try to avoid it by methods such as selecting extremely small positive numbers to replace or removing the indicators containing 0. The preparation link of data utilization also includes preparing the knowledge structure and quantity of experts. According to the field characteristics of the scientific research projects, appropriate technical experts and economic experts should be selected to balance their business capabilities, so as to achieve a comprehensive coverage of the capabilities of the evaluation indicators of the scientific research projects. At the same time, it is also necessary to ensure that the magnitudes of various scientific research inputs and scientific research outputs are as close as possible.

[0070] Step 3: Classify and determine the index system. Classify the scientific research projects into trial production type, technology type and research type, design and determine the corresponding input indicators and output indicators. The innovation output is divided into special indicators and general indicators, and corresponding selection is made according to actual application needs. The classification index system is constructed as shown in Table 1.

[0071] Table 1 Innovation Performance Index System

[0072]

[0073]

[0074] Score and assign values to different scientific research project DMO. To ensure the interpretability of the subsequent analysis results, an appropriate quantization unit should be selected to ensure that the input and output data are in integer form. This link is the key link affecting the effectiveness of the method of this patent, and it is necessary to design the format and accurately measure the data of each input and output indicator.

[0075] Step 4: First-stage efficiency evaluation. According to whether it is necessary to eliminate the influence of scientific research and production scale, the applicable models in the first-stage efficiency evaluation link are respectively the hybrid efficiency evaluation model or the pure efficiency evaluation model. The specific rule is that when the scale influence difference is small, the hybrid efficiency evaluation model is adopted; otherwise, the pure efficiency evaluation model is adopted. The basic connotations of the hybrid efficiency evaluation model and the pure efficiency evaluation model are elaborated as follows.

[0076] ① Hybrid efficiency evaluation model

[0077] Suppose we want to measure the first-stage efficiency of a group of n DMOs in total, denoted as DMO j (j = 1, 2, … n) in sequence; the input types of each DMO are m, and the output types are q. Their respective quantity values are denoted as x i (i = 1, 2, … m) and y r (r = 1, 2, … q) respectively. The input weights and output weights are denoted as v i (i = 1, 2, … m) and u r (r = 1, 2, … q) respectively. The target DMO to be measured is denoted as DMO k , and its output-input ratio is expressed as:

[0078]

[0079] Based on this, the equation of the hybrid efficiency evaluation model is given as:

[0080]

[0081] Perform model transformation. Let

[0082]

[0083] μ = tu,

[0084] γ = tv,

[0085] Transform the original model (3) into an equivalent linear programming model. The model form is as follows:

[0086]

[0087] The model (6) is called the multiplier form of hybrid efficiency measurement. Through dual transformation, the final expression of the hybrid efficiency evaluation model can be obtained as:

[0088]

[0089] In expression (7), λ represents the linear combination coefficient of DMO, and the optimal solution θ * of the model represents the efficiency value and satisfies the condition θ * ∈(0, 1]. 1 - θ *The economic meaning represented is the degree of inefficiency, θ * The smaller it is, the greater the scope for substantial reduction in input factors, the higher the degree of inefficiency, and the lower the efficiency value. When θ * = 1, it indicates that the k-th decision-making object DMO being evaluated k is on the innovation frontier surface. Under the condition of unchanged output, there is no room for proportional reduction in various input factors, and it is in an innovation-efficient state. When θ * < 1, it is collectively referred to as the innovation-inefficient state, and the maximum scope for proportional reduction in various input factors is calculated according to the ratio of 1 - θ * .

[0090] ② Pure efficiency evaluation model

[0091] The pure efficiency evaluation model expands the basic applicable conditions of the mixed efficiency evaluation model. When the scale of scientific and technological innovation changes and the impact degree is large and cannot be ignored, it is necessary to supplement and correct the mixed efficiency evaluation model to establish a pure efficiency evaluation model. The basic equation of the pure efficiency evaluation model is:

[0092]

[0093] In formula (8), the constraint condition is supplemented and λ ≥ 0. It can be understood that its role is to make the scale of scientific research input factors of the projection point at the same level as the scale of the DMO being evaluated. Similarly to the mixed efficiency evaluation model, the dual model expression of the pure efficiency evaluation model is:

[0094]

[0095] When examining the innovation efficiency of scientific research projects and there is no need to eliminate the influence of scale efficiency, the mixed efficiency evaluation model should be selected for efficiency evaluation; when examining the innovation efficiency of scientific research projects and it is necessary to eliminate the influence of scale efficiency, the pure efficiency evaluation model should be selected for efficiency evaluation. The relationship among the three is:

[0096] SES = CTRS / PTRS (10)

[0097] In the above formula, SES (Scale Efficiency Score) represents the scale efficiency value, CTRS (Comprehensive Technical Efficiency Score) represents the mixed innovation efficiency value, and PTRS (Pure Technical Efficiency Score) represents the pure innovation efficiency value.

[0098] Step 5: Second-order efficiency evaluation. When the hybrid efficiency evaluation model or the pure efficiency evaluation model is selected for evaluation, and the innovation efficiency values ​​of the DMOs of multiple scientific research projects are all 1, it is necessary to enter the second-order efficiency model evaluation link to improve the efficiency evaluation differentiation. By establishing a second-order efficiency evaluation model, further efficiency evaluation conclusions can be obtained, and scientific research projects can be ranked according to their efficiency. The basic connotation of the second-order efficiency evaluation model is explained as follows.

[0099] ③ Second-order efficiency evaluation model

[0100] The essence of the second-order efficiency evaluation model is that the reference object of the evaluated DMO is exclusive. By examining the innovation frontier formed by other decision-making objects, it is concluded that the effective DMO innovation efficiency value is usually greater than 1, thereby making a secondary distinction between effective decision-making objects.

[0101] Based on the basic assumption that the input-oriented scale efficiency remains unchanged, the constraint condition j≠k is added to transform the hybrid efficiency evaluation model into a second-order efficiency evaluation model. k For , it is necessary to remove it from the reference set, so the basic equation of the second-order efficiency evaluation model is:

[0102]

[0103] When input-oriented scale efficiency becomes the basic assumption, additional constraints The pure efficiency evaluation model is transformed into a second-order efficiency evaluation model. The basic equation of the second-order efficiency evaluation model is:

[0104]

[0105] Regardless of model (11) or (12), the second-order efficiency evaluation model is not only applicable to the innovation efficiency judgment of effective DMOs, but also to the innovation efficiency judgment of ineffective DMOs.

[0106] Step 6: Determine the performance of scientific research projects. Based on the innovation efficiency value and ranking conclusion of DMO, the maximum performance value of scientific research projects under the established rules (which can be a percentage system, a thousandth system, etc.) is given, and the two are multiplied to calculate the innovation performance value of all evaluated scientific research projects.

[0107] Example 2

[0108] Taking December 31, 2023 as the node, a total of 30 research projects that completed their research tasks before the node date were selected as decision-making objects (DMOs) to quantify and rank the sample innovation performance evaluation. For the convenience of explanation, the previous data collection, collation and calculation process is omitted, and the specific input and output indicators of the 30 DMOs are directly given, as shown in Table 2.

[0109] Table 2 Original data of “input-output” of decision-making object of innovation performance evaluation (DMO=30)

[0110]

[0111]

[0112]

[0113] The basic information contained in the above table is as follows: ① The number of scientific research project samples participating in the innovation performance evaluation is 30, that is, a scientific research performance evaluation model containing 30 decision-making objects is established; ② The scientific research input indicators are composed of 3 components, namely "scientific research funds", "scientific research time" and "scientific research personnel", and the units are ten thousand yuan, month and number of people respectively; ③ The scientific research output indicators are composed of 4 specific indicators and 2 general indicators. The specific indicators are "academic papers", "patents and software", and "monographs and reports", and the units are the number of articles, the number of items and the number of parts respectively; the general indicators are "number of application service personnel" and "number of people satisfied with scientific research improvement", and the units are 100 people and 100 people respectively; ④ All original indicators in the model are absolute quantity indicators.

[0114] Among them, the "value of research results" indicator needs to be calculated according to certain rules. It is usually calculated by combining weighting and expert evaluation according to the value of the results. Table 3 gives a detailed calculation rule for a commonly used indicator equivalent score. The specific indicator data in Table 2 is correlated with the data in Table 3.

[0115] Table 3 Calculation rules and scores for the value equivalent of “research results value” in innovation performance evaluation

[0116]

[0117]

[0118] In the above table, the three indicators of "academic papers", "patents and software", and "monographs and reports" are evaluated according to different value standards. Here, the specific classification level connotation is no longer defined, and the main purpose is to display the method. At this point, after the equivalent score is calculated according to the established rules, the complete panel data for this patent technology method is obtained, and the data preparation work is now complete.

[0119] When evaluating the efficiency of scientific research projects, different applicable models need to be selected according to different basic assumptions. When the condition of constant returns to scale is met, we choose the hybrid efficiency evaluation model to calculate and rank the efficiency of 30 scientific research project samples. The input settings of the hybrid efficiency evaluation model are as follows: decision-making object DMO = 30, input variables = 3, output variables = 4, and the innovation efficiency value and the determination result of the effectiveness status (ineffective, weakly effective, strongly effective) are obtained. Similarly, when the condition of variable returns to scale is met, we choose the pure efficiency evaluation model to eliminate the influence of scale efficiency. By analogy with the input setting values of the hybrid efficiency evaluation model, the innovation efficiency value and the effectiveness status of each project can also be obtained. The calculation results of the two are compared, and the relevant results are shown in Table 4 as follows.

[0120] Table 4 Comparison of the innovation efficiency of scientific research projects under two types of hypothesis conditions

[0121]

[0122]

[0123] Obviously, a total of 14 scientific research projects with an effective status are obtained by the calculation of the hybrid efficiency evaluation model, and all of them are in the "strongly effective" status. The results show that when the returns to scale are constant, the decision-making objects P1, P2, P3, P6, P7, P8, P12, P13, P14, P15, P17, P23, P25, P29 are technically effective, and the rest of the decision-making objects are in a technically ineffective state; similarly, a total of 23 scientific research projects with an "effective" status are obtained by the calculation of the pure efficiency evaluation model. Except that P9 is in the "weakly effective" status, the rest are in the strongly effective status. This result shows that when the returns to scale are variable, the decision-making objects P4, P5, P9, P11, P18, P19, P20, P26, P27 have changed from the original "ineffective" status to the "effective" status.

[0124] The comparison of the two results shows that when the influence factor of scale efficiency is eliminated, the obtained pure efficiency value shows a certain upward trend. According to formula (10), the influence degree of the calculated scale efficiency can be quantitatively reflected in Table 5.

[0125] Table 5 Comparison of the influence of scale efficiency

[0126]

[0127] Obviously, for the innovation efficiency values of each decision-making object obtained by the preliminary judgment, their discrimination degree still cannot reach the ideal standard. Taking the hybrid efficiency evaluation model as an example, the efficiency values of up to 14 decision-making objects are all 1. By introducing the second-order efficiency evaluation model, the innovation efficiency of the decision-making objects is judged twice, so as to obtain the final innovation efficiency value, performance value and ranking result, as shown in Table 6.

[0128] Table 6 List of performance values and rankings of scientific research projects based on the second-order efficiency evaluation model

[0129]

[0130]

[0131] According to the table, the performance values of 30 scientific research projects are sorted from largest to smallest, serving as the basis for subsequent innovation incentive distribution. The project ranked first in performance is P29, with a performance value as high as 187.1 points, and the project ranked last in performance is P9, with a performance value of only 44.2 points. In addition, there are a total of 16 projects with room for improvement in innovation efficiency. When the scientific research output remains unchanged, it is necessary to reduce the quantity of scientific research input factors, and the specific adjustment plan is shown in Table 7.

[0132] Table 7 Implementation plan for improving decision-making objects based on input orientation

[0133]

[0134]

[0135] The above conclusions fully demonstrate that the stepwise efficiency optimization method proposed in this application can not only objectively and comprehensively quantify and rank the innovation performance levels of scientific research projects, but also give specific improvement and optimization plans, specifying the quantity changes in the allocation of each input factor, and has good application prospects.

[0136] Example 3

[0137] As Figure 3 shown, in another embodiment of the present invention, a stepwise efficiency optimization innovation performance evaluation system is provided for implementing the stepwise efficiency optimization innovation performance evaluation method, including:

[0138] A data acquisition module for obtaining data to be evaluated, processing the data, eliminating negative sample data and data with zero indicators, and obtaining decision-making objects;

[0139] A data processing module for designing a classification index system according to the decision-making objects, scoring and assigning values to the sub-indicators, and performing normalization processing on the assignments;

[0140] A module establishment module for establishing a hybrid efficiency evaluation model, a pure efficiency evaluation model, and a second-order efficiency evaluation model, where:

[0141] When examining the innovation efficiency of scientific research projects without eliminating the influence of scale efficiency, efficiency evaluation is performed based on the hybrid efficiency evaluation model;

[0142] When examining the innovation efficiency of a scientific research project and it is necessary to eliminate the influence of scale efficiency, efficiency evaluation is carried out based on the pure efficiency evaluation model;

[0143] When there is a situation where the innovation efficiency values of multiple decision-making objects are all 1 in the evaluation results of the hybrid efficiency evaluation model or the pure efficiency evaluation model, evaluation is carried out based on the second-order efficiency evaluation model.

[0144] Those skilled in the art can easily understand that the above is only a preferred embodiment of the present invention and is not used to limit the present invention. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of the present invention shall be included within the protection scope of the present invention.

Claims

1. A step-by-step efficiency optimization innovation performance evaluation method, characterized in that: include: S100: Obtain the data to be evaluated, process the data, remove negative sample data and data containing zero indicators, and obtain a decision object; S200: designing a classification index system according to the decision object, assigning scores to the subdivided indexes, and normalizing the assigned scores; S300: Establishing a hybrid efficiency evaluation model, a pure efficiency evaluation model and a second-order efficiency evaluation model, wherein: When examining the innovation efficiency of scientific research projects and without eliminating the impact of scale efficiency, efficiency evaluation is performed based on the hybrid efficiency evaluation model; When examining the innovation efficiency of scientific research projects and eliminating the impact of scale efficiency, efficiency evaluation is performed based on the pure efficiency evaluation model; When the innovation efficiency values ​​of multiple decision-making objects are all 1 in the evaluation results of the hybrid efficiency evaluation model or the pure efficiency evaluation model, the evaluation is performed based on the second-order efficiency evaluation model.

2. The innovation performance evaluation method of step-by-step efficiency optimization according to claim 1 is characterized in that: In step S300, the establishment of the mixing efficiency evaluation model includes: S301: Determine a group of n decision objects, where the target decision object to be measured is DMO k , respectively denoted as DMO j (j=1,2,…n), its output-input ratio is expressed as: S302: Based on this, the mixed efficiency evaluation model is obtained as follows: The input type of each decision object is m, the output type is q, and their respective quantity values ​​are x i (i=1,2,…m) and y r (r=1,2,…q), the input weight and output weight are v i (i=1,2,…m) and u r (r=1,2,…q);k≤n.

3. The innovation performance evaluation method of step-by-step efficiency optimization according to claim 2 is characterized in that: In step S300, the establishment of the mixing efficiency evaluation model includes: transforming the mixing efficiency evaluation model, so that: μ=tu, γ=tv, The hybrid efficiency evaluation model is transformed into an equivalent linear programming model:

4. The innovation performance evaluation method of step-by-step efficiency optimization according to claim 3 is characterized in that: In step S300, the establishment of the hybrid efficiency evaluation model includes: performing dual transformation on the linear programming model to obtain the final expression of the hybrid efficiency evaluation model: In the formula, λ represents the linear combination coefficient of the decision object, and the optimal solution of the model is θ * Represents the efficiency value and satisfies the condition θ * ∈(0,1].

5. The innovation performance evaluation method of step-by-step efficiency optimization according to claim 4 is characterized in that: In step S300, the establishment of the pure efficiency evaluation model includes: S303: When the scale of technological innovation changes and the impact is too large to be ignored, it is necessary to supplement and revise the hybrid efficiency evaluation model and add constraints. The pure efficiency determination model is established as: In the formula, and λ≥0.

6. The innovation performance evaluation method of step-by-step efficiency optimization according to claim 5 is characterized in that: In step S300, the establishment of the pure efficiency evaluation model includes: performing a dual transformation on the pure efficiency determination model to obtain a final expression of the pure efficiency determination model:

7. The innovation performance evaluation method of step-by-step efficiency optimization according to claim 6 is characterized in that: In step S300, the establishment of the second-order efficiency evaluation model includes: based on the basic assumption that the input-oriented scale efficiency remains unchanged, supplementing the constraint condition j≠k, and transforming the hybrid efficiency evaluation model into a second-order efficiency evaluation model:

8. The innovation performance evaluation method of step-by-step efficiency optimization according to claim 7 is characterized in that: In step S300, the establishment of the second-order efficiency evaluation model includes: when the scale efficiency based on input orientation becomes the basic assumption, the constraint condition is supplemented The pure efficiency evaluation model is transformed into a second-order efficiency evaluation model:

9. A step-by-step efficiency optimization innovation performance evaluation method according to any one of claims 1 to 8, characterized in that: In step S100, the decision objects satisfy the following conditions: the number n of decision objects is greater than or equal to three times the number of input and output indicators, and the number n of decision objects is greater than or equal to the product of the number of input and output indicators, that is: n≥max{3×(m+q),m×q} Among them, m and q are the number of scientific research input indicators and the number of scientific research output indicators respectively.

10. A step-by-step efficiency optimization innovation performance evaluation system, characterized in that: An innovative performance evaluation method for realizing step-by-step efficiency optimization as claimed in any one of claims 1 to 9, comprising: The data acquisition module is used to obtain the data to be evaluated, process the data, remove negative sample data and data containing zero indicators, and obtain the decision object; A data processing module is used to design a classification index system according to the decision object, assign scores to the subdivided indicators, and normalize the assigned scores; The module establishment module is used to establish a mixed efficiency evaluation model, a pure efficiency evaluation model and a second-order efficiency evaluation model, where: When examining the innovation efficiency of scientific research projects and without eliminating the impact of scale efficiency, efficiency evaluation is performed based on the hybrid efficiency evaluation model; When examining the innovation efficiency of scientific research projects and eliminating the impact of scale efficiency, efficiency evaluation is performed based on the pure efficiency evaluation model; When the innovation efficiency values ​​of multiple decision-making objects are all 1 in the evaluation results of the hybrid efficiency evaluation model or the pure efficiency evaluation model, the evaluation is performed based on the second-order efficiency evaluation model.

Images