Science and technology project review optimization method and system based on artificial intelligence
Through artificial intelligence technology, the features are extracted from the scientific and technological project database, the text similarity is calculated and the fuzzy evaluation is performed, the results of scientific and technological project review are optimized, the problems of low efficiency and strong subjectivity of the traditional review system are solved, and more efficient and objective Review results.
Patent Information
- Application Number
- CN202411083792.3
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-08
- Publication Date
- 2024-11-26
AI Technical Summary
The existing scientific and technological project review system adopts the traditional manual review method, which has slow processing speed and low efficiency, lacks objective review standards and technical support. The review results are easily affected by personal experience and subjective judgments, and it is difficult to ensure fairness and accuracy, and lacks Intelligent support cannot effectively process large amounts of data and information.
Using the scientific and technological project evaluation optimization method based on artificial intelligence, the characteristics that affect the success rate of the project are extracted from the scientific and technological project database, the text similarity is calculated using the cosine similarity method, and the fuzzy set transformation principle and the fuzzy evaluation matrix are combined to determine the Review optimization factors and optimize review results.
Improve the efficiency and accuracy of the review, assist experts in making decisions, correct experts’ subjective judgments, improve the objectivity and fairness of the review, and provide intelligent review opinions management functions, help project leaders track and improve review opinions, and improve Review quality and reliability.
Smart Images

Figure CN119027282A_ABST
Abstract
Description
Technical Field
[0001] The present invention belongs to the field of artificial intelligence technology, and more specifically, relates to an artificial intelligence-based technology project review optimization method and system. Background Art
[0002] Science and technology project review is the process of evaluating and making decisions on science and technology projects. The review process is usually conducted by experts or a review committee to ensure that the project complies with the science and technology development plan and industrial policy. At the same time, it evaluates the project's funding, budget rationality, and the rationality of project demand investment. It is of great significance to improve the efficiency of the use of science and technology funds, promote scientific and technological innovation and industrial development.
[0003] The existing science and technology project review process mainly includes first determining the project review objectives and standards based on the science and technology development plan and industrial policy, then publicly soliciting science and technology projects from the society, and reviewing and screening the accepted projects to ensure the quality and implementation value of the projects. Then, a preliminary review of the accepted projects is conducted, and the project leaders are organized to defend the projects. The review committee asks questions and evaluates the projects on site. Finally, based on the review standards and the defense situation, the projects are comprehensively evaluated and demonstrated to form a review opinion.
[0004] However, there are still some shortcomings in its actual use. For example, the existing science and technology project review system usually adopts the traditional manual review method, which has slow processing speed, low efficiency, lack of objective review standards and technical support, and the review results are often affected by personal experience and subjective judgment. It is difficult to ensure the fairness and accuracy of the review. There is a lack of effective quality supervision mechanism, and it is difficult to ensure the stability and reliability of the review quality. There is a lack of intelligent support, and it is impossible to automatically analyze and process large amounts of data and information, making it difficult to provide accurate decision support. Summary of the invention
[0005] In order to solve the deficiencies in the prior art, the present invention provides a scientific and technological project review optimization system based on artificial intelligence.
[0006] The present invention adopts the following technical solution.
[0007] The first aspect of the present invention provides a method for optimizing scientific and technological project evaluation based on artificial intelligence, comprising the following steps:
[0008] Extract the features that affect the project success rate from the information obtained from the scientific and technological project database to establish a feature set;
[0009] Based on the established feature set, the cosine similarity method is used to calculate the similarity of the scientific and technological project texts;
[0010] Based on the obtained feature set and the calculated text similarity, the fuzzy set transformation principle and fuzzy evaluation matrix are used to conduct comprehensive evaluation and simulated review of scientific and technological projects, and determine the review optimization factor of each scientific and technological project;
[0011] Use review optimization factors to optimize the review results of scientific and technological projects.
[0012] Preferably, extracting features that affect project success rate from information obtained from the scientific and technological project database to establish a feature set includes:
[0013] According to the review content and review form, it is classified into management review and technical review. Management review is used to identify and solve management problems and make management decision reviews, while technical review is used to identify and solve technical problems and make technical decision reviews.
[0014] The information of scientific and technological projects is collected through the scientific and technological project database, including project name, field, technical characteristics, R&D team, investment scale and completion status, and marked in sequence as λ{λ1, λ2, …, λ m}, m represents the number of scientific and technological project information;
[0015] Based on the collected science and technology project information, the common features β related to science and technology project evaluation are extracted, including technology maturity, market demand, team strength and investment scale, and the differentiated features τ are extracted, including technology novelty, application prospects and business model;
[0016] Based on the extracted common features and differentiated features, a common feature set λ is established β {λ β1 ,λ β2 , …, λ βm} and the differentiated feature set λ τ {λ τ1 ,λ τ2 , …, λ τm}, each feature is standardized.
[0017] Preferably, the method of calculating the similarity of the scientific and technological project texts based on the established feature set using the cosine similarity method includes:
[0018] The basic similarity measure is used to compare the directions of vectors and is as follows:
[0019]
[0020] Where:
[0021] Attribute vector A and attribute vector B represent the word frequency vectors in the scientific and technological project text;
[0022] A·B is specifically The representation shows the frequency of two texts on the same lexical dimension;
[0023] ||A||||B||Specifically Indicates the amount of vocabulary used in the text;
[0024] cos(θ) represents the similarity between two vectors, 1 means they are exactly the same, 0 means they are irrelevant, and -1 means they are completely opposite.
[0025] Preferably, the method of calculating the similarity of the scientific and technological project texts based on the established feature set using the cosine similarity method also includes:
[0026] Calculate the arccosine of the basic similarity measure cos -1 (similarity), used to convert the similarity between two vectors into an angle value;
[0027] Taking into account the signs of the vector coefficients, choose the appropriate normalization formula:
[0028] If the vector coefficient is positive or negative, use the following formula:
[0029]
[0030] If the vector coefficients are always positive, use the following formula:
[0031]
[0032] Substitute the calculated angle value into the corresponding normalization formula to obtain a bounded similarity value in the range [0,1].
[0033] Preferably, the feature set based on step 1 and the text similarity calculated in step 2 is used to perform comprehensive evaluation and simulated review on the scientific and technological projects using the fuzzy set transformation principle and the fuzzy judgment matrix to determine the review optimization factors of each scientific and technological project. By determining the review factors, the review grade standards of the factors and the weights, the fuzzy set transformation principle is used to describe the fuzzy boundaries of each factor and the factor with the degree of membership, and a fuzzy judgment matrix is constructed. Through composite operations, comprehensive evaluation and simulated review are performed to determine the grade of the reviewed scientific and technological projects, including:
[0034] Determine the evaluation factors U = {U1, U2, ..., U N},U i Represents the review factors, where i = 1, 2, ..., N, where N represents the number of individual factors at the same level, forming the framework of the review;
[0035] Determine the review level standard set V = {V1, V2, ..., V n}, V = (1, 2, ..., n) represents the evaluation grade standard, n represents the number of elements, specifically the number of grades or the number of comment files, V = (1, 2, ..., n) specifies the selection range of the evaluation results of a single evaluation factor, and the evaluation element is a qualitative or quantitative score;
[0036] Conduct a separate review of each review factor, construct a fuzzy evaluation matrix, and record the degree of membership of each factor relative to each review grade standard;
[0037] Customize review criteria and weights through the user interface, calculate and display review results in real time, provide comparison and visualization tools, analyze and select the best review solution;
[0038] According to the membership degree in the fuzzy evaluation matrix, the maximum membership principle is adopted to determine the final review level of the project as the review optimization factor.
[0039] Preferably, the user interface is used to customize the review criteria and weights, calculate and display the review results in real time, provide comparison and visualization tools, and analyze and select the optimal review solution, including:
[0040] For each factor U in the review factor set i , and conduct a separate review to obtain j The fuzzy vector R j =(r i1 , r i2 ,…,r ij ), where R j =(r i1 , r i2 ,…,r ij ), j = 1, 2, ..., n, r ij Factor U i With V j The degree of, and 0≤r ij ≤1;
[0041] Comprehensively review n elements to obtain a matrix with N rows and n columns, which represents the membership degree R. Each row in the matrix is the review result of each single factor.
[0042] The membership of qualitative indicators is obtained through expert scoring method, and the membership of quantitative indicators is calculated by applying membership function to generate a comment set;
[0043] The fuzzy vector R corresponding to all review factors U i The membership matrix R is combined. The number of rows in the membership matrix represents the number of evaluation factors N, and the number of columns represents the number of evaluation level standards n. Each element in the matrix represents the membership of the evaluation factor relative to the corresponding evaluation level.
[0044] Preferably, the user interface is used to customize the review criteria and weights, calculate and display the review results in real time, provide comparison and visualization tools, and analyze and select the optimal review solution, including:
[0045] Provide a user interface to customize the review criteria and the weight of each review factor according to project characteristics and review requirements;
[0046] Based on the set review criteria and weights, the review results of each project are calculated in real time, and the review grades and conclusions under different review criteria and weights are displayed;
[0047] Provide comparison tools to conduct comparative analysis on multiple review plans, display the review results and success rate prediction of projects under each plan, and select the best plan;
[0048] Provides visualization tools, including comparison charts, scoring tables, and review reports, to intuitively display the results of different review schemes and export simulated review results in text or table format for analysis and recording.
[0049] Preferably, the optimization of the scientific and technological project review results by using the review optimization factor includes:
[0050] Based on the obtained review optimization factors, the project leader and relevant researchers screen and search the review opinions according to the classification labels of the science and technology projects or the artificial conditions proposed by the science and technology projects;
[0051] Set sharing options, and the project leader will share the review opinions with specific users and set the permissions for viewing and discussing;
[0052] Set reminder and notification rules, specify who needs to be notified when review comments are added, modified, or discussed, and select notification methods, including email, system notification, and mobile push;
[0053] Automatically monitor the optimization factors of review opinions, trigger notification rules, generate notifications based on changes, and send them to relevant personnel.
[0054] The second aspect of the present invention provides a scientific and technological project review optimization system based on artificial intelligence, comprising: a feature extraction module, a text similarity calculation module, a comprehensive evaluation module and a review optimization module;
[0055] The feature extraction module is used to extract the features that affect the project success rate from the information obtained from the scientific and technological project database and establish a feature set;
[0056] The text similarity calculation module is used to calculate the scientific and technological project text using the cosine similarity method to obtain the text similarity;
[0057] The comprehensive evaluation module is used to use the fuzzy set transformation principle and fuzzy evaluation matrix to conduct comprehensive evaluation and simulated review of scientific and technological projects and determine the evaluation optimization factors of each scientific and technological project;
[0058] The review optimization module is used to optimize the review results of scientific and technological projects using review optimization factors.
[0059] The third aspect of the present invention provides a terminal, including a processor and a storage medium; the storage medium is used to store instructions; the processor is used to operate according to the instructions to execute the steps of a technology project review optimization method based on artificial intelligence.
[0060] The fourth aspect of the present invention provides a computer-readable storage medium having a computer program stored thereon, characterized in that the program is executed by a processor to perform the steps of a scientific and technological project review optimization method based on artificial intelligence.
[0061] Compared with the prior art, the beneficial effects of the present invention include at least:
[0062] Through artificial intelligence technology, the present invention can quickly and accurately process a large amount of review data and conduct multi-dimensional screening, thereby improving review efficiency, assisting experts in making decisions, avoiding simple repetitive work and wasting human resources, and at the same time checking for deficiencies and filling in the gaps, helping experts in multi-dimensional screening, and improving the objectivity and accuracy of the review. Through intelligent detection results and text matching analysis, it can assist experts in making decisions, correct the objectivity of the review of supervisory experts, and improve the efficiency of expert analysis. At the same time, it provides review opinion management functions to help project leaders or relevant researchers better organize and track review opinions, and provide strong support for the improvement and development of scientific and technological projects.
[0063] (1) The present invention improves data processing efficiency and accuracy through automated data extraction and feature analysis, ensuring the comprehensiveness and scientificity of data in the review process, thereby improving the objectivity and accuracy of the review results;
[0064] (2) The present invention uses artificial intelligence technology to quickly and accurately process a large amount of review data and conduct multi-dimensional screening, thereby improving review efficiency, assisting experts in making decisions, avoiding simple repetitive work and wasting human resources, and at the same time checking for deficiencies and filling gaps, helping experts in multi-dimensional screening, and improving the objectivity and accuracy of the review; through intelligent detection results and text matching analysis, it assists experts in making decisions, corrects the objectivity of the review of supervisory experts, and improves the efficiency of expert analysis. At the same time, it provides review opinion management functions to help project leaders or relevant researchers better organize and track review opinions, and provide strong support for the improvement and development of scientific and technological projects;
[0065] (3) The present invention realizes scientific and reasonable comprehensive evaluation by applying the fuzzy set transformation principle and fuzzy evaluation matrix, provides a user interface for customizing evaluation criteria and weights, and improves the flexibility, scientificity and accuracy of evaluation and decision support through real-time calculation, visualization tools and comparative analysis;
[0066] (4) The present invention provides enhanced query, screening and modification functions through intelligent review opinion management, and sets reminder and notification functions to ensure the timeliness and interactivity of the review process, effectively improve the quality and fairness of the review, and ensure the reliability of the review results. BRIEF DESCRIPTION OF THE DRAWINGS
[0067] Figure 1 It is an overall structural diagram provided according to an embodiment of the present invention. DETAILED DESCRIPTION
[0068] In order to make the purpose, technical scheme and advantages of the present invention clearer, the technical scheme of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. The embodiments described in this application are only part of the embodiments of the present invention, not all of them. Based on the spirit of the present invention, other embodiments obtained by ordinary technicians in this field without creative work are all within the scope of protection of the present invention.
[0069] like Figure 1 As shown, embodiment 1 of the present invention provides a method for optimizing scientific and technological project review based on artificial intelligence, comprising the following steps:
[0070] Extract the features that affect the project success rate from the information obtained from the scientific and technological project database to establish a feature set;
[0071] Preferably, extracting features that affect project success rate from information obtained from the scientific and technological project database to establish a feature set includes:
[0072] According to the review content and review form, it is classified into management review and technical review. Management review is used to identify and solve management problems and make management decision reviews, while technical review is used to identify and solve technical problems and make technical decision reviews.
[0073] The information of scientific and technological projects is collected through the scientific and technological project database, including project name, field, technical characteristics, R&D team, investment scale and completion status, and marked in sequence as λ{λ1, λ2, …, λ m}, m represents the number of scientific and technological project information;
[0074] Based on the collected science and technology project information, the common features β related to science and technology project evaluation are extracted, including technology maturity, market demand, team strength and investment scale, and the differentiated features τ are extracted, including technology novelty, application prospects and business model;
[0075] Based on the extracted common features and differentiated features, a common feature set λ is established β {λ β1 ,λ β2 , …, λ βm} and the differentiated feature set λ τ {λ τ1 ,λ τ2 , …, λ τm}, each feature is standardized.
[0076] Based on the established feature set, the cosine similarity method is used to match the scientific and technological project texts, calculate the similarity of the texts, and normalize the similarity of the project texts;
[0077] Preferably, the method of matching the scientific and technological project texts based on the established feature set using the cosine similarity method, calculating the similarity of the texts, and normalizing the similarity of the project texts includes:
[0078] Compute the cosine between two vectors using the Euclidean dot product:
[0079] a·b=||a||||b||cosθ;
[0080] Establish attribute vector A and attribute vector B, calculate basic similarity measures, and standardize file lengths when comparing texts;
[0081] Preferably, the attribute vector A and the attribute vector B are established, and a basic similarity metric is calculated. In the process of comparing texts and normalizing the file length, the basic similarity metric is used to compare the direction of the vectors, and the formula is as follows:
[0082]
[0083] Where:
[0084] Attribute vector A and attribute vector B represent the word frequency vectors in the scientific and technological project text;
[0085] A·b is specifically The representation shows the frequency of two texts on the same lexical dimension;
[0086] ||A||||B||Specifically Indicates the amount of vocabulary used in the text;
[0087] cos(θ) represents the similarity between two vectors, 1 means they are exactly the same, 0 means they are irrelevant, and -1 means they are completely opposite.
[0088] Based on the obtained basic similarity measure, normalization processing is performed to measure and compare text similarity.
[0089] Preferably, the normalization process based on the obtained basic similarity measurement, measuring and comparing text similarity includes:
[0090] Calculate the arccosine of the basic similarity measure cos -1 (similarity), used to convert the similarity between two vectors into an angle value;
[0091] Taking into account the signs of the vector coefficients, choose the appropriate normalization formula:
[0092] If the vector coefficient is positive or negative, use the following formula:
[0093]
[0094] If the vector coefficients are always positive, use the following formula:
[0095]
[0096] Substitute the calculated angle value into the corresponding normalization formula to obtain a bounded similarity value in the range [0,1].
[0097] Based on the obtained feature set and the calculated text similarity, the fuzzy set transformation principle and fuzzy evaluation matrix are used to conduct comprehensive evaluation and simulated review of scientific and technological projects, and determine the review optimization factor of each scientific and technological project;
[0098] Specifically, the obtained feature set and the calculated text similarity are used to perform comprehensive evaluation and simulated review of the scientific and technological projects using the fuzzy set transformation principle and the fuzzy judgment matrix, and the review optimization factors of each scientific and technological project are determined. On the basis of determining the review factors, the review grade standards of the factors and the weights, the fuzzy set transformation principle is used to describe the fuzzy boundaries of each factor and the factor with the degree of membership, and the fuzzy judgment matrix is constructed. Through composite operations, comprehensive evaluation and simulated review are performed to determine the grade of the reviewed scientific and technological projects, including:
[0099] Determine the evaluation factors U = {U1, U2, ..., U N},U i Represents the review factors, where i = 1, 2, ..., N, where N represents the number of individual factors at the same level, forming the framework of the review;
[0100] Determine the review level standard set V = {V1, V2, ..., V n}, V = (1, 2, ..., n) represents the evaluation grade standard, n represents the number of elements, specifically the number of grades or the number of comment files, V = (1, 2, ..., n) specifies the selection range of the evaluation results of a single evaluation factor, and the evaluation element is a qualitative or quantitative score;
[0101] Conduct a separate review of each review factor, construct a fuzzy evaluation matrix, and record the degree of membership of each factor relative to each review grade standard;
[0102] Preferably, the step of individually reviewing each review factor, constructing a fuzzy evaluation matrix, and recording the degree of membership of each factor relative to each review grade standard includes:
[0103] For each factor U in the review factor set i , and conduct a separate review to obtain j The fuzzy vector R j =(r i1 , r i2 ,…,r ij ), where R j =(r i1 , r i2 ,…,r ij ), j = 1, 2, ..., n, r ij Factor U i With V j The degree of, and 0≤r ij ≤1;
[0104] Comprehensively review n elements to obtain a matrix with N rows and n columns, which represents the membership degree R. Each row in the matrix is the review result of each single factor.
[0105] The membership of qualitative indicators is obtained through expert scoring method, and the membership of quantitative indicators is calculated by applying membership function to generate a comment set;
[0106] The fuzzy vector R corresponding to all review factors U i Combined into a membership matrix R, the number of rows of the membership matrix represents the number of evaluation factors N, the number of columns represents the number of evaluation level standards n, and each element in the matrix represents the membership of the evaluation factor relative to the corresponding evaluation level.
[0107] Customize review criteria and weights through the user interface, calculate and display review results in real time, provide comparison and visualization tools, analyze and select the best review solution;
[0108] Preferably, the user interface is used to customize the review criteria and weights, calculate and display the review results in real time, provide comparison and visualization tools, and analyze and select the optimal review solution, including:
[0109] Provide a user interface to customize the review criteria and the weight of each review factor according to project characteristics and review requirements;
[0110] Based on the set review criteria and weights, the review results of each project are calculated in real time, and the review grades and conclusions under different review criteria and weights are displayed;
[0111] Through the comparison tool, multiple review plans are compared and analyzed, the review results and success rate predictions of the projects under each plan are displayed, and the best plan is selected;
[0112] Through visual tools, including comparison charts, score sheets, and review reports, the results of different review schemes can be intuitively displayed, and the simulated review results can be exported as text or tables for analysis and recording.
[0113] According to the membership degree in the fuzzy evaluation matrix, the maximum membership principle is adopted to determine the final review level of the project as the review optimization factor.
[0114] Use review optimization factors to optimize the review results of scientific and technological projects, manage and screen review opinions, modify review opinions, and assist in decision-making project reviews.
[0115] Preferably, the review optimization factor is used as a key parameter to further optimize the review results of scientific and technological projects. The specific process is as follows:
[0116] Integrate the obtained review project grade data into the review system to form a standardized data set, including the characteristics, text similarity and review grade of each project;
[0117] Using review optimization factors, we establish a review result optimization model. By combining machine learning algorithms and big data analysis technology, we can conduct deep learning and pattern recognition based on a large amount of historical data and existing project data, thereby identifying key factors and potential problems that affect review results.
[0118] In the actual review process, the characteristics and level changes of the review items are monitored in real time, and the parameters of the optimization model are continuously adjusted to ensure the accuracy and effectiveness of the review results. The optimized review results will be immediately fed back to the review committee to assist it in making the final decision;
[0119] Based on the analysis results of the review optimization factors, it is recommended to adjust the review standards and weights to meet the review needs of different types of projects. For example, for projects with strong technological innovation, the weight of technology maturity can be appropriately increased; for projects with clear market orientation, the review weight of market demand and business model can be increased.
[0120] Preferably, the use of review optimization factors to optimize the review results of scientific and technological projects, manage and screen review opinions, modify review opinions, and assist in decision-making project review includes:
[0121] Based on the generated review optimization factors, the project leader and relevant researchers screen and search for review opinions according to the classification labels of science and technology projects or artificial conditions proposed by science and technology projects;
[0122] Set sharing options, and the project leader will share the review opinions with specific users and set the permissions for viewing and discussing;
[0123] Set reminder and notification rules, specify who needs to be notified when review comments are added, modified, or discussed, and select notification methods, including email, system notification, and mobile push;
[0124] Automatically monitor changes in review optimization factors, trigger notification rules, generate notifications based on changes, and send them to relevant personnel.
[0125] Example 2 of the present invention provides a scientific and technological project review optimization system based on artificial intelligence, including: a feature extraction module, a text similarity calculation module, a comprehensive evaluation module and a review optimization module;
[0126] The feature extraction module is used to extract the features that affect the project success rate from the information obtained from the scientific and technological project database and establish a feature set;
[0127] The text similarity calculation module is used to calculate the scientific and technological project text using the cosine similarity method to obtain the text similarity;
[0128] The comprehensive evaluation module is used to use the fuzzy set transformation principle and fuzzy evaluation matrix to conduct comprehensive evaluation and simulated review of scientific and technological projects and determine the evaluation optimization factors of each scientific and technological project;
[0129] The review optimization module is used to optimize the review results of scientific and technological projects using review optimization factors.
[0130] The present invention establishes a common feature set and a differentiated feature set from the original data feature set of the scientific and technological project through a feature selection module, obtains the similarity of two texts in a statistical method through a text matching module, applies the fuzzy set transformation principle through a comprehensive evaluation module, describes the fuzzy boundaries of each factor and the factor with the membership degree, constructs a fuzzy evaluation matrix, and finally determines the grade of the scientific and technological project to be reviewed through composite operations. The review opinion management module provides query and screening functions for the review opinions, supports exporting the review opinions in the form of text or table, and sets review opinion reminder and notification functions.
[0131] The beneficial effects of the present invention are as follows:
[0132] (1) The present invention improves data processing efficiency and accuracy through automated data extraction and feature analysis, ensuring the comprehensiveness and scientificity of data in the review process, thereby improving the objectivity and accuracy of the review results;
[0133] (2) The present invention uses artificial intelligence technology to quickly and accurately process a large amount of review data and conduct multi-dimensional screening, thereby improving review efficiency, assisting experts in making decisions, avoiding simple repetitive work and wasting human resources, and at the same time checking for deficiencies and filling gaps, helping experts in multi-dimensional screening, and improving the objectivity and accuracy of the review; through intelligent detection results and text matching analysis, it assists experts in making decisions, corrects the objectivity of the review of supervisory experts, and improves the efficiency of expert analysis. At the same time, it provides review opinion management functions to help project leaders or relevant researchers better organize and track review opinions, and provide strong support for the improvement and development of scientific and technological projects;
[0134] (3) The present invention realizes scientific and reasonable comprehensive evaluation by applying the fuzzy set transformation principle and fuzzy evaluation matrix, provides a user interface for customizing evaluation criteria and weights, and improves the flexibility, scientificity and accuracy of evaluation and decision support through real-time calculation, visualization tools and comparative analysis;
[0135] (4) The present invention provides enhanced query, screening and modification functions through intelligent review opinion management, and sets reminder and notification functions to ensure the timeliness and interactivity of the review process, effectively improve the quality and fairness of the review, and ensure the reliability of the review results.
[0136] The present disclosure may be a system, a method and / or a computer program product. The computer program product may include a computer-readable storage medium carrying computer-readable program instructions for causing a processor to implement various aspects of the present disclosure.
[0137] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the above embodiments, ordinary technicians in the relevant field should understand that the specific implementation methods of the present invention can still be modified or replaced by equivalents, and any modifications or equivalent replacements that do not depart from the spirit and scope of the present invention should be covered within the scope of protection of the claims of the present invention.
Claims
1. A scientific and technological project review optimization method based on artificial intelligence, characterized in that: The following steps are involved: Extract the features that affect the project success rate from the information obtained from the scientific and technological project database to establish a feature set; Based on the established feature set, the cosine similarity method is used to calculate the similarity of the scientific and technological project texts; Based on the obtained feature set and the calculated text similarity, the fuzzy set transformation principle and fuzzy evaluation matrix are used to conduct comprehensive evaluation and simulated review of scientific and technological projects, and determine the review optimization factor of each scientific and technological project; Use review optimization factors to optimize the review results of scientific and technological projects.
2. According to the artificial intelligence-based scientific and technological project evaluation optimization method of claim 1, it is characterized by: The feature set is established by extracting features that affect the project success rate from the information obtained from the scientific and technological project database, including: According to the review content and review form, it is classified into management review and technical review. Management review is used to identify and solve management problems and make management decision reviews, while technical review is used to identify and solve technical problems and make technical decision reviews. The information of scientific and technological projects is collected through the scientific and technological project database, including project name, field, technical characteristics, R&D team, investment scale and completion status, and marked in sequence as λ{λ1, λ2, …, λ m }, m represents the number of scientific and technological project information; Based on the collected science and technology project information, the common features β related to science and technology project evaluation are extracted, including technology maturity, market demand, team strength and investment scale, and the differentiated features τ are extracted, including technology novelty, application prospects and business model; Based on the extracted common features and differentiated features, a common feature set λ is established β {λ β1 ,λ β2 , …, λ βm } and the differentiated feature set λ τ {λ τ1 ,λ τ2 , …, λ τm }, each feature is standardized.
3. According to the artificial intelligence-based scientific and technological project evaluation optimization method of claim 1, it is characterized by: The method of calculating the similarity of the scientific and technological project texts based on the established feature set using the cosine similarity method includes: The basic similarity measure is used to compare the directions of vectors and is as follows: Where: Attribute vector A and attribute vector B represent the word frequency vectors in the scientific and technological project text; A·B is specifically The representation shows the frequency of two texts on the same lexical dimension; ||A||||B||Specifically Indicates the amount of vocabulary used in the text; cos(θ) represents the similarity between two vectors, 1 means they are exactly the same, 0 means they are irrelevant, and -1 means they are completely opposite.
4. The method for optimizing scientific and technological project evaluation based on artificial intelligence according to claim 1, characterized in that: The method of calculating the similarity of the scientific and technological project texts based on the established feature set using the cosine similarity method also includes: Calculate the arccosine of the basic similarity measure cos -1 (similarity), used to convert the similarity between two vectors into an angle value; Taking into account the signs of the vector coefficients, choose the appropriate normalization formula: If the vector coefficient is positive or negative, use the following formula: If the vector coefficients are always positive, use the following formula: Substitute the calculated angle value into the corresponding normalization formula to obtain a bounded similarity value in the range [0,1].
5. The method for optimizing scientific and technological project evaluation based on artificial intelligence according to claim 1 is characterized in that: Based on the obtained feature set and the calculated text similarity, the fuzzy set transformation principle and the fuzzy judgment matrix are used to conduct comprehensive evaluation and simulated review of the scientific and technological projects, and the review optimization factors of each scientific and technological project are determined. On the basis of determining the review factors, the review grade standards of the factors and the weights, the fuzzy set transformation principle is used to describe the fuzzy boundaries of each factor and the factors with the membership degree, and the fuzzy judgment matrix is constructed. Through composite operations, comprehensive evaluation and simulated review are conducted to determine the grade of the reviewed scientific and technological projects, including: Determine the evaluation factors U = {U1, U2, ..., U N },U i Represents the review factors, where i = 1, 2, ..., N, where N represents the number of individual factors at the same level, forming the framework of the review; Determine the review level standard set V = {V1, V2, ..., V n }, V = (1, 2, ..., n) represents the evaluation grade standard, n represents the number of elements, specifically the number of grades or the number of comment files, V = (1, 2, ..., n) specifies the selection range of the evaluation results of a single evaluation factor, and the evaluation element is a qualitative or quantitative score; Conduct a separate review of each review factor, construct a fuzzy evaluation matrix, and record the degree of membership of each factor relative to each review grade standard; Customize review criteria and weights through the user interface, calculate and display review results in real time, provide comparison and visualization tools, analyze and select the best review solution; According to the membership degree in the fuzzy evaluation matrix, the maximum membership principle is adopted to determine the final review level of the project as the review optimization factor.
6. The method for optimizing scientific and technological project evaluation based on artificial intelligence according to claim 5 is characterized in that: The user interface is used to customize review criteria and weights, calculate and display review results in real time, provide comparison and visualization tools, and analyze and select the best review solution, including: For each factor U in the review factor set i , and conduct a separate review to obtain j The fuzzy vector R j =(r i1 , r i2 ,…,r ij ), where R j =(r i1 , r i2 ,…,r ij ), j = 1, 2, ..., n, r ij Factor U i With V j The degree of, and 0≤r ij ≤1; Comprehensively review n elements to obtain a matrix with N rows and n columns, which represents the membership degree R. Each row in the matrix is the review result of each single factor. The membership of qualitative indicators is obtained through expert scoring method, and the membership of quantitative indicators is calculated by applying membership function to generate a comment set; The fuzzy vector R corresponding to all review factors U i The membership matrix R is combined. The number of rows in the membership matrix represents the number of evaluation factors N, and the number of columns represents the number of evaluation level standards n. Each element in the matrix represents the membership of the evaluation factor relative to the corresponding evaluation level.
7. The method for optimizing scientific and technological project evaluation based on artificial intelligence according to claim 5 is characterized in that: The user interface is used to customize review criteria and weights, calculate and display review results in real time, provide comparison and visualization tools, and analyze and select the best review solution, including: Provide a user interface to customize the review criteria and the weight of each review factor according to project characteristics and review requirements; Based on the set review criteria and weights, the review results of each project are calculated in real time, and the review grades and conclusions under different review criteria and weights are displayed; Provide comparison tools to conduct comparative analysis on multiple review plans, display the review results and success rate prediction of projects under each plan, and select the best plan; Provides visualization tools, including comparison charts, scoring tables, and review reports, to intuitively display the results of different review schemes and export simulated review results in text or table format for analysis and recording.
8. The method for optimizing scientific and technological project evaluation based on artificial intelligence according to claim 1 is characterized in that: The optimization of the scientific and technological project review results by using the review optimization factors includes: Based on the obtained review optimization factors, the project leader and relevant researchers screen and search the review opinions according to the classification labels of the science and technology projects or the artificial conditions proposed by the science and technology projects; Set sharing options, and the project leader will share the review opinions with specific users and set the permissions for viewing and discussing; Set reminder and notification rules, specify who needs to be notified when review comments are added, modified, or discussed, and select notification methods, including email, system notification, and mobile push; Automatically monitor the optimization factors of review opinions, trigger notification rules, generate notifications based on changes, and send them to relevant personnel.
9. A scientific and technological project review optimization system based on artificial intelligence, comprising: Feature extraction module, text similarity calculation module, comprehensive evaluation module and review optimization module; Running an artificial intelligence-based scientific and technological project review optimization method as described in any one of claims 1 to 8, characterized in that: The feature extraction module is used to extract the features that affect the project success rate from the information obtained from the scientific and technological project database and establish a feature set; The text similarity calculation module is used to calculate the scientific and technological project text using the cosine similarity method to obtain the text similarity; The comprehensive evaluation module is used to use the fuzzy set transformation principle and fuzzy evaluation matrix to conduct comprehensive evaluation and simulated review of scientific and technological projects and determine the evaluation optimization factors of each scientific and technological project; The review optimization module is used to optimize the review results of scientific and technological projects using review optimization factors.
10. A terminal comprising a processor and a storage medium; characterized in that: The storage medium is used to store instructions; The processor is used to operate according to the instructions to execute the steps of the artificial intelligence-based science and technology project review optimization method according to any one of claims 1-8.
11. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the steps of an artificial intelligence-based technology project evaluation optimization method described in any one of claims 1 to 8 are implemented.
Citation Information
Cited By
Intelligent project review method and system based on data analysis
CN119963000A
Scientific research project question-making review information optimization method and scientific research project question-making review information optimization device
CN121255956A
A scientific research project completion review information optimization method and device
CN121255956B
