Innovation and entrepreneurship practice project intelligent evaluation system based on AI driving

Abstract

The invention relates to the field of project evaluation, in particular to an AI-driven intelligent evaluation system for innovation and entrepreneurship practice projects, which comprises the following steps: extracting information keywords corresponding to practice projects, performing retrieval in a project database, determining the content overlap ratio with a plurality of projects and the keyword overlap number, and constructing a fuzzy matching project set; identifying a farthest matching time difference between the practice item and a plurality of fuzzy matching items, calibrating a farthest matching time interval, calculating an innovation fuzzy characterization value of the practice item based on fuzzy matching features, dividing innovation degree categories of the practice item, adaptively evaluating and analyzing the practice item, and determining the innovation degree of the practice item. Comprising the following steps: based on specific keywords of a practice item relative to a plurality of fuzzy matching items in a fuzzy matching item set, carrying out joint evaluation on the specific keywords; and determining a fuzzy innovation level of the practice item according to the innovation fuzzy representation value, and storing the fuzzy innovation level in a database. According to the invention, the precision and efficiency of practical project evaluation are improved.

Description

Technical Field

[0001] This invention relates to the field of project evaluation, and in particular to an AI-driven intelligent evaluation system for innovation and entrepreneurship practice projects. Background Technology

[0002] In the current innovation and entrepreneurship ecosystem, the complexity and diversity of innovation and entrepreneurship practice projects continue to increase, further complicating the evaluation work. On the one hand, there is a significant increase in cross-disciplinary integration projects, requiring evaluators to possess multidisciplinary knowledge backgrounds in order to accurately judge the rationality and innovativeness of technological integration; on the other hand, the forms of project innovation are becoming increasingly diverse. In addition to traditional technological breakthrough innovations, new forms of innovation such as model innovation, scenario innovation, and service innovation are constantly emerging, placing higher demands on the comprehensiveness and flexibility of evaluation standards.

[0003] In addition, the shorter project lifecycle and faster iteration speed also require the evaluation work to be more timely, able to quickly respond to the needs of project development, and provide timely support for the phased adjustments and resource allocation of the project.

[0004] Against this backdrop, traditional project evaluation models are no longer sufficient to meet the development needs of current innovation and entrepreneurship practices. There is an urgent need to leverage the advantages of artificial intelligence technology to build an intelligent and systematic evaluation system to improve the efficiency, accuracy, and objectivity of evaluations, better serve the decision-making needs of various stakeholders, and promote the healthy development of the innovation and entrepreneurship ecosystem.

[0005] Chinese Patent Application Publication No. CN118917723A discloses a comprehensive innovation and entrepreneurship capability assessment system and method. The system includes a comprehensive capability data acquisition module, a comprehensive capability data processing module, a comprehensive capability data analysis module, a comprehensive capability evaluation coefficient calculation module, a comprehensive capability assessment module, and an assessment result feedback module. The system collects comprehensive capability data through the comprehensive capability data acquisition module, processes the comprehensive capability data through the comprehensive capability data processing module, analyzes the secondary feature data through the comprehensive capability data analysis module, calculates the comprehensive capability evaluation coefficient through the comprehensive capability evaluation coefficient calculation module, assesses the comprehensive innovation and entrepreneurship capability based on the comprehensive capability evaluation coefficient, and provides feedback on the assessment results to individuals or organizations through the assessment result feedback module. By processing and analyzing a large amount of comprehensive capability data from multiple perspectives, the assessment results are made more accurate and reliable.

[0006] However, the following problems still exist in the existing technology. Traditional assessments often rely on qualitative judgments based on expert experience, lacking quantitative indicators and unified standards. This can easily lead to inconsistent assessment results due to differences in expert background and cognition, thus reducing the accuracy of the assessment. Summary of the Invention

[0007] To address this, the present invention provides an AI-driven intelligent evaluation system for innovation and entrepreneurship practice projects, which overcomes the problems of existing technologies that rely heavily on qualitative judgments based on expert experience, lack quantitative indicators and unified standards, and are prone to inconsistent evaluation results due to differences in expert background and cognition, thus reducing the accuracy of the evaluation.

[0008] To achieve the above objectives, this invention provides an AI-driven intelligent evaluation system for innovation and entrepreneurship practice projects, comprising: The project retrieval module is used to extract information keywords corresponding to practical projects, search the project database, determine the content overlap and keyword overlap with several projects, and construct a fuzzy matching project set. An innovation assessment module, connected to the project retrieval module, is used to identify the furthest matching time difference between the practice project and several fuzzy matching projects, and to mark the corresponding furthest matching time interval, determine the fuzzy matching features corresponding to the furthest matching time interval, and calculate the innovation fuzzy representation value of the practice project to classify the innovation degree category of the practice project. The distinguishing module is connected to the innovation evaluation module and, in response to the division result of the innovation evaluation module, selects to call either the fuzzy evaluation module or the clear evaluation module; A clear evaluation module, connected to the distinguishing call module, is used to jointly evaluate the special keywords based on the special keywords of the practice item relative to several fuzzy matching items in the fuzzy matching item set. This includes evaluating the novelty characterization parameters of the practice item based on the novelty distinguishing features of the special keywords, so as to perform derivation verification on the practice item. Identify the derivative words of the specific keywords, and determine whether the derivative words meet the derivative expansion criteria based on the expansion of applicable scenarios and the degree of interpretation difference of the specific keywords, so as to determine whether the practice project is marked as an innovative project and stored in the database; A fuzzy evaluation module, connected to the distinguishing and calling module, is used to determine the fuzzy innovation level of the practice project based on the innovation fuzzy characterization value and store it in the database; The fuzzy matching features include the number of fuzzy matching items and the average increase in content overlap, while the novel distinguishing features include the highest semantic similarity and the number of overlapping applicable scenarios.

[0009] Furthermore, the item retrieval module is used to construct a fuzzy matching item set, including: Used to call up information keywords for practical projects and project keywords for several projects; If any item meets the fuzzy matching criteria, then the item is identified as a fuzzy matching item. The set of several fuzzy matching items is used to define the fuzzy matching item set. The fuzzy matching conditions include a content overlap degree greater than a content overlap degree threshold and a keyword overlap number greater than a keyword overlap number threshold.

[0010] Furthermore, the innovation evaluation module is used to calibrate the corresponding furthest matching time interval, including: This is used to match the project initiation time of the practice project with the publication time of several fuzzy matching projects; Used to determine the furthest publication time in the time dimension relative to the project initiation time; This is used to determine the furthest matching time interval by taking the project initiation time as the upper limit and the furthest publication time as the lower limit. The furthest matching time interval is a closed interval.

[0011] Furthermore, the innovation evaluation module is used to calculate the innovation fuzzy representation value of the practice project, including: The ratio of the furthest matching time difference threshold to the furthest matching time difference is used as the first innovative fuzzy feature; The sum of the ratio of the number of fuzzy matching items to the threshold of the number of fuzzy matching items and the ratio of the average increase in content overlap to the threshold of the average increase in content overlap is used as the second innovative fuzzy feature. The first innovation fuzzy feature and the second innovation fuzzy feature are weighted and summed to determine the innovation fuzzy representation value.

[0012] Furthermore, the innovation assessment module is used to categorize the degree of innovation of the practice projects, including: If the innovation fuzzy representation value of the practice project is greater than or equal to the innovation fuzzy representation threshold, then the innovation level category of the practice project will be classified as low innovation level category; If the innovation fuzzy representation value of a practice project is less than the innovation fuzzy representation threshold, then the innovation level category of the practice project will be classified as a high innovation level category.

[0013] Furthermore, the distinguishing invocation module is used to select whether to invoke the fuzzy evaluation module or the succinct evaluation module, including: If the innovation level category of the practice project is low, then the fuzzy evaluation module is invoked; If the innovation level of the practice project is classified as high innovation, then the clear assessment module will be invoked.

[0014] Furthermore, the clarity assessment module is used to evaluate the novelty characterization parameters of the practice project, including: The ratio of the highest semantic similarity to the highest semantic similarity threshold is used as the first novelty feature; The ratio of the number of overlapping applicable scenarios to the threshold number of overlapping applicable scenarios is used as a second novelty feature; The sum of the first novelty feature and the second novelty feature is used as the novelty characterization parameter.

[0015] Furthermore, the clear evaluation module is used to determine whether the derived extension benchmark is met, including: If the expansion of the applicable scenarios of the derived words is less than the threshold for the expansion of applicable scenarios, and the degree of explanation difference of the special keywords is less than the threshold for explanation difference, then the derivative expansion benchmark is met.

[0016] Furthermore, the clear assessment module is used to determine whether the practice project should be marked as an innovative project, including: If a practice project meets the derived extension criteria, then the practice project is marked as an innovative project.

[0017] Furthermore, the fuzzy evaluation module is used to determine the fuzzy innovation level of the practice project based on the innovation fuzzy representation value, including: Pre-set the correspondence between fuzzy innovation levels and predetermined fuzzy characterization value ranges; Determine the range of innovation fuzzy representation values ​​to which the corresponding innovation fuzzy representation values ​​of the practice projects belong; The fuzzy innovation level corresponding to the fuzzy characterization value range of the innovation is determined as the fuzzy innovation level of the practical project; Among them, the fuzzy innovation level corresponds one-to-one with the range of fuzzy innovation representation values.

[0018] Compared with existing technologies, this invention includes a project retrieval module, which extracts information keywords corresponding to practical projects, searches a project database, determines the content overlap and keyword overlap with several projects, and constructs a fuzzy matching project set; an innovation evaluation module, which identifies the furthest matching time difference between the practical project and several fuzzy matching projects, marks the corresponding furthest matching time interval, determines the fuzzy matching features corresponding to the furthest matching time interval, calculates the innovation fuzzy representation value of the practical project, and classifies the innovation degree category of the practical project; and a differentiation and invocation module, which, in response to the classification result of the innovation evaluation module, selects to invoke either the fuzzy evaluation module or the clear evaluation module to evaluate and analyze the practical project. This invention, while adapting to the diversity and dynamism of innovation and entrepreneurship practical projects, improves the accuracy and efficiency of practical project evaluation.

[0019] In particular, this invention sets up an innovation evaluation module, locking in the innovation reference area of ​​practical projects in the time dimension, i.e., the furthest matching time difference. It is not solely limited to content similarity but also identifies corresponding time intervals, integrating the time dimension with content similarity. This makes the subsequent judgment of the innovation level of practical projects more aligned with the time context of their development. By extracting fuzzy matching features, the similarity between practical projects and similar projects is quantified. The number of fuzzy matching projects reflects the density of similar projects corresponding to the practical project, quantifying the scale of homogenization in the field. Furthermore, it focuses on the degree of overlap of the core content of the practical project with each fuzzy matching project, eliminating random deviations of individual projects through the average increase in content overlap, and quantifying the depth of similarity and the degree of differentiation of the core content of the practical projects. Therefore, this invention calculates an innovation fuzzy representation value for practical projects to characterize the comprehensive degree of difference between the practical project and existing similar projects, as well as the degree of innovation of the practical project, providing data support for subsequent classification of the innovation level of practical projects. This invention improves the accuracy and efficiency of practical project evaluation while adapting to the diversity and dynamism of innovation and entrepreneurship practical projects.

[0020] In particular, this invention establishes a clear evaluation module to extract unique keywords that distinguish practical projects from fuzzy-matched projects, i.e., existing similar projects, filtering out homogeneous information interference and focusing on the innovative elements of practical projects. The highest semantic similarity measure quantifies the difference between the unique keywords of the practical project and the project keywords of the fuzzy-matched project, reflecting its closeness to similar terms in existing publicly available information; the number of overlapping applicable scenarios reflects the overlap range of application scenarios between the practical project and the fuzzy-matched project. Therefore, this invention characterizes the novelty of practical projects through novelty-level parameters and conducts derivation verification based on this, forming a closed-loop logic of quantitative evaluation and in-depth verification, thus improving the accuracy and efficiency of practical project evaluation.

[0021] In particular, when conducting in-depth verification of practical projects, this invention considers the derived terms of unique elements within the projects to assess their extensibility, avoiding the misjudgment of projects that merely replace keywords without substantial innovation as innovative projects. The expansion of applicable scenarios characterizes the actual coverage and extension of application scenarios for each project. Simultaneously, the degree of interpretive difference reflects the difference between the project and existing lexical interpretations, essentially judging whether the innovative concept of the project represents a substantial breakthrough, and characterizing the uniqueness of the innovative conceptual connotation. Therefore, a comprehensive determination is made as to whether a project is innovative. This invention, while adapting to the diversity and dynamism of innovative and entrepreneurial practical projects, improves the accuracy and efficiency of project evaluation. Attached Figure Description

[0022] Figure 1A functional block diagram of an AI-driven intelligent evaluation system for innovation and entrepreneurship practice projects, as an embodiment of the invention; Figure 2 A logic diagram for classifying the degree of innovation of practical projects in the embodiments of the invention; Figure 3 A logic decision diagram for selecting whether to call the fuzzy evaluation module or the clear evaluation module in the embodiments of the invention; Figure 4 A logic decision diagram for determining whether a derived extension benchmark is satisfied in an embodiment of the invention. Detailed Implementation

[0023] To make the objectives and advantages of the present invention clearer, the present invention will be further described below with reference to embodiments; it should be understood that the specific embodiments described herein are merely for explaining the present invention and are not intended to limit the present invention.

[0024] Preferred embodiments of the present invention will now be described with reference to the accompanying drawings. Those skilled in the art should understand that these embodiments are merely illustrative of the technical principles of the present invention and are not intended to limit the scope of protection of the present invention.

[0025] It should be noted that in the description of this invention, the terms "upper," "lower," "inner," etc., indicating the direction or positional relationship are based on the direction or positional relationship shown in the drawings. This is only for the convenience of description and is not intended to indicate or imply that the device or element must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, it should not be construed as a limitation of this invention.

[0026] Furthermore, it should be noted that, in the description of this invention, unless otherwise explicitly specified and limited, the term "connection" should be interpreted broadly. For example, it can refer to a fixed connection, a detachable connection, or an integral connection; it can refer to a mechanical connection or an electrical connection. Those skilled in the art can understand the specific meaning of the above terms in this invention according to the specific circumstances.

[0027] Please see Figure 1 The diagram shown is a functional block diagram of an AI-driven intelligent evaluation system for innovation and entrepreneurship practice projects according to an embodiment of the present invention. The AI-driven intelligent evaluation system for innovation and entrepreneurship practice projects according to an embodiment of the present invention includes: The project retrieval module is used to extract information keywords corresponding to practical projects, search the project database, determine the content overlap and keyword overlap with several projects, and construct a fuzzy matching project set. An innovation assessment module, connected to the project retrieval module, is used to identify the furthest matching time difference between the practice project and several fuzzy matching projects, and to mark the corresponding furthest matching time interval, determine the fuzzy matching features corresponding to the furthest matching time interval, and calculate the innovation fuzzy representation value of the practice project to classify the innovation degree category of the practice project. The distinguishing module is connected to the innovation evaluation module and, in response to the division result of the innovation evaluation module, selects to call either the fuzzy evaluation module or the clear evaluation module; A clear evaluation module, connected to the distinguishing call module, is used to jointly evaluate the special keywords based on the special keywords of the practice item relative to several fuzzy matching items in the fuzzy matching item set. This includes evaluating the novelty characterization parameters of the practice item based on the novelty distinguishing features of the special keywords, so as to perform derivation verification on the practice item. Identify the derivative words of the specific keywords, and determine whether the derivative words meet the derivative expansion criteria based on the expansion of applicable scenarios and the degree of interpretation difference of the specific keywords, so as to determine whether the practice project is marked as an innovative project and stored in the database; A fuzzy evaluation module, connected to the distinguishing and calling module, is used to determine the fuzzy innovation level of the practice project based on the innovation fuzzy characterization value and store it in the database; The fuzzy matching features include the number of fuzzy matching items and the average increase in content overlap, while the novel distinguishing features include the highest semantic similarity and the number of overlapping applicable scenarios.

[0028] Specifically, the project database refers to a database that stores publicly available project data collected from public channels, such as innovation and entrepreneurship competition project databases, as well as cooperative channels, such as university innovation and entrepreneurship centers. This will not be elaborated further.

[0029] Specifically, there are no restrictions on the specific structure of the project retrieval module, innovation evaluation module, differentiation and call module, clear evaluation module, and fuzzy evaluation module. Each module or its units can be composed of logical components or combinations of logical components. Logical components include field-programmable processors, computers, or microprocessors in computers.

[0030] Specifically, there are no specific limitations on how to determine the content overlap between the practical project and several projects. The semantic overlap of the overall content between projects can be determined by a pre-trained language model. Of course, other methods can also be used to determine the content overlap, which will not be elaborated here.

[0031] Specifically, the method for determining the semantic similarity between a specific keyword and the corresponding project keywords in a fuzzy matching project can be any of the existing technologies, such as word vectors. Then, the highest semantic similarity is determined, which will not be elaborated further.

[0032] Specifically, the item retrieval module is used to construct a fuzzy matching item set, including: Used to call up information keywords for practical projects and project keywords for several projects; If any item meets the fuzzy matching criteria, then the item is identified as a fuzzy matching item. The set of several fuzzy matching items is used to define the fuzzy matching item set. The fuzzy matching conditions include a content overlap degree greater than a content overlap degree threshold and a keyword overlap number greater than a keyword overlap number threshold.

[0033] Specifically, based on the information keywords of practical projects and the project keywords of several projects, special keywords are determined, and information keywords that are different from the keywords of each project are used as the special keywords. Specifically, the cosine similarity between information keywords and project keywords can be calculated to determine whether there is a distinguishing relationship between them. Therefore, in practice, information keywords with a cosine similarity of less than 0.3 are used as the specific keywords.

[0034] In this embodiment, the purpose of setting the content overlap threshold and the keyword overlap threshold is to characterize the situation where the information keywords of the practice project have a high degree of overlap with the content of other projects. Therefore, based on the purpose of setting the above two thresholds, the content overlap threshold is set to 60%, and the keyword overlap threshold is set to 65%.

[0035] Specifically, the innovation evaluation module is used to calibrate the corresponding furthest matching time interval, including: This is used to match the project initiation time of the practice project with the publication time of several fuzzy matching projects; Used to determine the furthest publication time in the time dimension relative to the project initiation time; This is used to determine the furthest matching time interval by taking the project initiation time as the upper limit and the furthest publication time as the lower limit. The furthest matching time interval is a closed interval.

[0036] Specifically, this invention determines the furthest matching time interval based on the time dimension, which can quickly respond to the characteristics of shortened project life cycle and accelerated iteration speed, timely capture the innovative value of the time dimension of practical projects, provide timely evaluation support for project phase adjustments, and better serve the development needs of the innovation and entrepreneurship ecosystem.

[0037] Specifically, the innovation evaluation module is used to calculate the innovation fuzzy representation value of the practice project, including: The ratio of the furthest matching time difference threshold to the furthest matching time difference is used as the first innovative fuzzy feature; The sum of the ratio of the number of fuzzy matching items to the threshold of the number of fuzzy matching items and the ratio of the average increase in content overlap to the threshold of the average increase in content overlap is used as the second innovative fuzzy feature. The first innovation fuzzy feature and the second innovation fuzzy feature are weighted and summed to determine the innovation fuzzy representation value.

[0038] Specifically, the two features—the number of fuzzy matching projects and the average increase in content overlap—essentially quantify the degree of content homogenization between the practice project and existing similar projects from two dimensions: the scale of similar projects and the similarity of their core content. This allows for the analysis of whether the practice project possesses a differentiated advantage at the content level. The furthest matching time difference quantifies the time span between the practice project and existing similar projects. A closer time interval between the practice project and the furthest similar project indicates limited innovation space in the time dimension. The difference in the time dimension has an indirect and non-deterministic impact on the degree of innovation of the practice project. Therefore, the second innovative fuzzy feature calculated based on fuzzy matching features is assigned a higher weight coefficient, set to 0.6, while the weight coefficient corresponding to the first innovative fuzzy feature calculated based on the furthest matching time difference is set to 0.4.

[0039] In this embodiment, the purpose of setting the threshold for the furthest matching time difference, the threshold for the number of fuzzy matching items, and the threshold for the average increase in content overlap is to characterize the situation where the content of the practice project and several fuzzy matching items overlaps significantly. By obtaining relevant data from the project database corresponding to several similar projects of the practice project, and calling the data on the furthest matching time difference between the practice project and several fuzzy matching items, the data on the number of fuzzy matching items within the corresponding furthest matching time interval, and the average increase in content overlap, the average furthest matching time difference, the average number of fuzzy matching items, and the average increase in content overlap are calculated and used as the baseline values ​​under normal circumstances. The purpose of the threshold is to determine the farthest matching time difference threshold as the product of the average farthest matching time difference and the first deviation coefficient, the fuzzy matching item quantity threshold as the product of the average fuzzy matching item quantity and the second deviation coefficient, and the content overlap average increase threshold as the product of the average content overlap average increase and the third deviation coefficient. The first deviation coefficient is selected within the interval [0.85, 0.9], preferably 0.85 in practice; the second deviation coefficient is selected within the interval [1.2, 1.4], preferably 1.2 in practice; and the third deviation coefficient is selected within the interval [1.15, 1.2], preferably 1.15 in practice.

[0040] Specifically, this invention sets up an innovation evaluation module to lock in the innovation reference area of ​​practical projects in the time dimension, that is, the furthest matching time difference. It is not limited to content similarity alone, but also marks the corresponding time interval, integrating the time dimension with content similarity, so that the subsequent judgment of the innovation degree of practical projects is more in line with the time context of the development of practical projects. By extracting fuzzy matching features, the similarity between practical projects and similar projects is quantified. The number of fuzzy matching projects reflects the density of similar projects corresponding to the practical project, quantifying the scale of homogenization in the field. If there are more fuzzy matching projects, it indicates that the core direction of the practical project has a higher coverage in existing practical projects, and the degree of homogenization in the field or technical direction is higher. Furthermore, it focuses on the degree of overlap of the core content of the practical project with each fuzzy matching project, and eliminates the random deviation of individual projects by the average increase in content overlap, quantifying the similarity depth and the degree of differentiation of the core content of the practical project. If the average increase in content overlap is larger, it indicates that the core content similarity and overlap of the practical project with similar projects are higher. Therefore, this invention calculates an innovation fuzzy representation value for a practical project to characterize the overall difference between the practical project and existing similar projects, as well as the degree of innovation of the practical project, providing data support for subsequent classification of the innovation degree of practical projects. This invention improves the accuracy and efficiency of practical project evaluation while adapting to the diversity and dynamism of innovation and entrepreneurship practical projects.

[0041] Specifically, please refer to Figure 2 As shown, this is a logical decision diagram for classifying the innovation level of practical projects according to an embodiment of the present invention. The innovation evaluation module is used to classify the innovation level of the practical projects, including: If the innovation fuzzy representation value of the practice project is greater than or equal to the innovation fuzzy representation threshold, then the innovation level category of the practice project will be classified as low innovation level category; If the innovation fuzzy representation value of a practice project is less than the innovation fuzzy representation threshold, then the innovation level category of the practice project will be classified as a high innovation level category.

[0042] The innovation fuzzy representation threshold C0 is predetermined. The innovation fuzzy representation value calculated under the following conditions is equal to the farthest matching time difference threshold, equal to the number of fuzzy matching items threshold, and equal to the average increase in content overlap threshold.

[0043] Specifically, please refer to Figure 3 As shown, this is a logic decision diagram for selecting to call the fuzzy evaluation module or the clear evaluation module in an embodiment of the present invention. The distinguishing call module is used to select to call the fuzzy evaluation module or the clear evaluation module, including: If the innovation level category of the practice project is low, then the fuzzy evaluation module is invoked; If the innovation level of the practice project is classified as high innovation, then the clear assessment module will be invoked.

[0044] Specifically, the clarity assessment module is used to evaluate the novelty characterization parameters of the practice project, including: The ratio of the highest semantic similarity to the highest semantic similarity threshold is used as the first novelty feature; The ratio of the number of overlapping applicable scenarios to the threshold number of overlapping applicable scenarios is used as a second novelty feature; The sum of the first novelty feature and the second novelty feature is used as the novelty characterization parameter.

[0045] In this embodiment, the purpose of setting the highest semantic similarity threshold and the applicable scenario overlap threshold is to characterize the low novelty of the practice project. By obtaining relevant data from the project database corresponding to several similar projects of the practice project, calling the highest semantic similarity data of specific keywords and the applicable scenario overlap data, the mean of the highest semantic similarity and the mean of the applicable scenario overlap are calculated, and the corresponding values ​​are used as the benchmark values ​​under normal circumstances. Based on the purpose of setting the above two thresholds, the highest semantic similarity threshold is determined as the product of the mean of the highest semantic similarity and the semantic deviation coefficient, and the applicable scenario overlap threshold is determined as the product of the mean of the applicable scenario overlap and the overlap deviation coefficient. The semantic deviation coefficient is selected in the interval [1.2, 1.25], preferably 1.2 in the implementation, and the overlap deviation coefficient is selected in the interval [1.2, 1.4], preferably 1.4 in the implementation.

[0046] Specifically, this invention establishes a clear evaluation module to extract unique keywords that distinguish practical projects from fuzzy matching projects, i.e., existing similar projects, filtering out homogeneous information interference and focusing on the innovative elements of practical projects. The highest semantic similarity measure quantifies the difference between the unique keywords of a practical project and the project keywords of a fuzzy matching project, reflecting its proximity to similar terms in existing public information. A lower highest semantic similarity indicates a greater semantic difference between the unique keywords and existing terms, and stronger novelty. The number of overlapping applicable scenarios reflects the overlap range of application scenarios between the practical project and the fuzzy matching project. A smaller number of overlapping applicable scenarios indicates a more unique application scenario corresponding to the unique keywords, and a more obvious differentiated advantage. Therefore, this invention uses novelty characterization parameters to represent the novelty of practical projects and conducts derivative verification based on this, forming a closed-loop logic of quantitative evaluation and in-depth verification, thus improving the accuracy and efficiency of practical project evaluation.

[0047] Specifically, the clear assessment module is used to perform derivative verification of the practice project, including: If the novelty characterization parameter of the practice project is less than the novelty characterization parameter threshold, then the practice project is subjected to derivation verification.

[0048] The novelty characterization parameter threshold is predetermined. The novelty characterization parameter threshold is determined by calculating the highest semantic similarity as equal to the highest semantic similarity threshold and the number of overlapping applicable scenarios as equal to the number of overlapping applicable scenarios threshold.

[0049] Specifically, please refer to Figure 4 As shown, this is a logic diagram for determining whether a derived extension benchmark is satisfied according to an embodiment of the present invention. The clear evaluation module is used to determine whether a derived extension benchmark is satisfied, including: If the expansion of the applicable scenarios of the derived words is less than the threshold for the expansion of applicable scenarios, and the degree of explanation difference of the special keywords is less than the threshold for explanation difference, then the derivative expansion benchmark is met.

[0050] Specifically, the degree of interpretation difference refers to the similarity between the explicit meaning description of specific keywords in practical projects and the meaning interpretations corresponding to keywords in several projects.

[0051] In this embodiment, the purpose of setting the applicable scenario expansion threshold and the explanation difference threshold is to characterize the situation where the practical project has a high degree of expansion but low novelty. By obtaining relevant data from the project database corresponding to several similar projects of the practical project, calling the applicable scenario expansion data of derived words and the explanation difference data of corresponding special keywords, the mean of applicable scenario expansion and the mean of explanation difference are calculated and used as the benchmark value under normal circumstances. Based on the purpose of setting the above two thresholds, the applicable scenario expansion threshold is determined as the product of the mean of applicable scenario expansion and the scenario deviation coefficient, and the explanation difference threshold is determined as the product of the explanation difference threshold and the explanation deviation coefficient. The scenario deviation coefficient is selected in the interval [0.9, 0.95], preferably 0.9 in the implementation, and the explanation deviation coefficient is selected in the interval [0.8, 0.9], preferably 0.8 in the implementation.

[0052] Specifically, the clear assessment module is used to determine whether the practice project should be marked as an innovative project, including: If a practice project meets the derived extension criteria, then the practice project is marked as an innovative project.

[0053] Specifically, when conducting in-depth verification of practical projects, this invention considers the derived terms of unique elements within the projects to assess their extensibility, avoiding the misjudgment of projects that merely replace keywords without substantial innovation as innovative projects. The scope of applicable scenarios is used to characterize the actual application scenario coverage and extension of the practical project. Simultaneously, the degree of interpretability reflects the difference between the practical project and existing lexical concepts, essentially judging whether the innovative concept of the practical project has a substantial breakthrough. It characterizes the uniqueness of the innovative conceptual connotation of the practical project. A high degree of interpretability indicates that the connotation of the unique keywords is more novel, not merely a slight modification or synonymous substitution of existing concepts, demonstrating a high degree of substantial conceptual innovation. Therefore, a comprehensive determination is made as to whether the practical project is innovative. This invention, while adapting to the diversity and dynamism of innovative and entrepreneurial practical projects, improves the accuracy and efficiency of practical project evaluation.

[0054] Specifically, the fuzzy evaluation module is used to determine the fuzzy innovation level of the practice project based on the innovation fuzzy characterization value, including: Pre-set the correspondence between fuzzy innovation levels and predetermined fuzzy characterization value ranges; Determine the range of innovation fuzzy representation values ​​to which the corresponding innovation fuzzy representation values ​​of the practice projects belong; The fuzzy innovation level corresponding to the fuzzy characterization value range of the innovation is determined as the fuzzy innovation level of the practical project; Among them, the fuzzy innovation level corresponds one-to-one with the range of fuzzy innovation representation values.

[0055] In this embodiment, the fuzzy innovation level of the practice project is determined in the following manner: The range of fuzzy representation values ​​related to innovation is divided into three preset ranges, and three levels of fuzzy innovation are set. If the innovation fuzzy representation value corresponding to the practice project is within the first preset interval (0, C0), then it corresponds to the first fuzzy innovation level; If the innovation fuzzy representation value corresponding to the practice project is within the second preset interval [C0, 1.3C0), then it corresponds to the second fuzzy innovation level; If the innovation fuzzy representation value corresponding to the practice project is within the third preset interval [1.3C0,+∞), then it corresponds to the third fuzzy innovation level.

[0056] Specifically, this invention stores practice projects marked as innovative projects in a database, realizing the systematic archiving of practice projects. This facilitates subsequent tracking and analysis of practice projects, such as summarizing innovation trends and reviewing value, and also provides data support for the traceability of evaluation results, thereby improving the professionalism and manageability of the evaluation system.

[0057] The technical solution of the present invention has been described above with reference to the preferred embodiments shown in the accompanying drawings. However, it will be readily understood by those skilled in the art that the scope of protection of the present invention is obviously not limited to these specific embodiments. Without departing from the principles of the present invention, those skilled in the art can make equivalent changes or substitutions to the relevant technical features, and the technical solutions after these changes or substitutions will all fall within the scope of protection of the present invention.

Claims

1. An AI-driven intelligent evaluation system for innovation and entrepreneurship practice projects, characterized in that, include: The project retrieval module is used to extract information keywords corresponding to practical projects, search the project database, determine the content overlap and keyword overlap with several projects, and construct a fuzzy matching project set. The innovation assessment module is used to identify the furthest matching time difference between the practice project and several fuzzy matching projects, mark the corresponding furthest matching time interval, determine the fuzzy matching features corresponding to the furthest matching time interval, and calculate the innovation fuzzy representation value of the practice project to classify the innovation degree category of the practice project. The module to be invoked is differentiated, and in response to the division result of the innovation evaluation module, it selects to invoke either the fuzzy evaluation module or the clear evaluation module. A clear evaluation module is used to jointly evaluate the specific keywords of the practice item relative to several fuzzy matching items in the fuzzy matching item set, including evaluating the novelty characterization parameters of the practice item based on the novelty distinguishing features of the specific keywords, so as to perform derivation verification of the practice item. Identify the derivative words of the specific keywords, and determine whether the derivative words meet the derivative expansion criteria based on the expansion of applicable scenarios and the degree of interpretation difference of the specific keywords, so as to determine whether the practice project is marked as an innovative project and stored in the database; A fuzzy evaluation module is used to determine the fuzzy innovation level of the practice project based on the innovation fuzzy characterization value and store it in the database; The fuzzy matching features include the number of fuzzy matching items and the average increase in content overlap, while the novel distinguishing features include the highest semantic similarity and the number of overlapping applicable scenarios.

2. The AI-driven intelligent evaluation system for innovation and entrepreneurship practice projects according to claim 1, characterized in that, The item retrieval module is used to construct a fuzzy matching item set, including: Used to call up information keywords for practical projects and project keywords for several projects; If any item meets the fuzzy matching criteria, then the item is identified as a fuzzy matching item. The set of several fuzzy matching items is used to define the fuzzy matching item set. The fuzzy matching conditions include a content overlap degree greater than a content overlap degree threshold and a keyword overlap number greater than a keyword overlap number threshold.

3. The AI-driven intelligent evaluation system for innovation and entrepreneurship practice projects according to claim 1, characterized in that, The innovation evaluation module is used to identify the corresponding furthest matching time interval, including: This is used to match the project initiation time of the practice project with the publication time of several fuzzy matching projects; Used to determine the furthest publication time in the time dimension relative to the project initiation time; This is used to determine the furthest matching time interval by taking the project initiation time as the upper limit and the furthest publication time as the lower limit. The furthest matching time interval is a closed interval.

4. The AI-driven intelligent evaluation system for innovation and entrepreneurship practice projects according to claim 1, characterized in that, The innovation assessment module is used to calculate the innovation fuzzy representation value of the practice project, including: The ratio of the furthest matching time difference threshold to the furthest matching time difference is used as the first innovative fuzzy feature; The sum of the ratio of the number of fuzzy matching items to the threshold of the number of fuzzy matching items and the ratio of the average increase in content overlap to the threshold of the average increase in content overlap is used as the second innovative fuzzy feature. The first innovation fuzzy feature and the second innovation fuzzy feature are weighted and summed to determine the innovation fuzzy representation value.

5. The AI-driven intelligent evaluation system for innovation and entrepreneurship practice projects according to claim 4, characterized in that, The innovation assessment module is used to classify the degree of innovation of the practice projects, including: If the innovation fuzzy representation value of the practice project is greater than or equal to the innovation fuzzy representation threshold, then the innovation level category of the practice project will be classified as low innovation level category; If the innovation fuzzy representation value of a practice project is less than the innovation fuzzy representation threshold, then the innovation level category of the practice project will be classified as a high innovation level category.

6. The AI-driven intelligent evaluation system for innovation and entrepreneurship practice projects according to claim 5, characterized in that, The distinguishing invocation module is used to select whether to invoke the fuzzy evaluation module or the sharp evaluation module, including: If the innovation level category of the practice project is low, then the fuzzy evaluation module is invoked; If the innovation level of the practice project is classified as high innovation, then the clear assessment module will be invoked.

7. The AI-driven intelligent evaluation system for innovation and entrepreneurship practice projects according to claim 1, characterized in that, The clarity assessment module is used to evaluate the novelty characterization parameters of the practice project, including: The ratio of the highest semantic similarity to the highest semantic similarity threshold is used as the first novelty feature; The ratio of the number of overlapping applicable scenarios to the threshold number of overlapping applicable scenarios is used as a second novelty feature; The sum of the first novelty feature and the second novelty feature is used as the novelty characterization parameter.

8. The AI-driven intelligent evaluation system for innovation and entrepreneurship practice projects according to claim 1, characterized in that, The clear evaluation module is used to determine whether the derived extension benchmark is met, including: If the expansion of the applicable scenarios of the derived words is less than the threshold for the expansion of applicable scenarios, and the degree of explanation difference of the special keywords is less than the threshold for explanation difference, then the derivative expansion benchmark is met.

9. The AI-driven intelligent evaluation system for innovation and entrepreneurship practice projects according to claim 8, characterized in that, The clear assessment module is used to determine whether the practice project should be marked as an innovative project, including: If a practice project meets the derived extension criteria, then the practice project is marked as an innovative project.

10. The AI-driven intelligent evaluation system for innovation and entrepreneurship practice projects according to claim 1, characterized in that, The fuzzy evaluation module is used to determine the fuzzy innovation level of the practice project based on the innovation fuzzy characterization value, including: Pre-set the correspondence between fuzzy innovation levels and predetermined fuzzy characterization value ranges; Determine the range of innovation fuzzy representation values ​​to which the corresponding innovation fuzzy representation values ​​of the practice projects belong; The fuzzy innovation level corresponding to the fuzzy characterization value range of the innovation is determined as the fuzzy innovation level of the practical project; Among them, the fuzzy innovation level corresponds one-to-one with the range of fuzzy innovation representation values.

Images