Project comprehensive evaluation method and system
By building a comprehensive project evaluation system based on convolutional prediction neural network, and automatically predicting project evaluation, the problems of high manual evaluation cost and deviation in result are solved, and efficient and accurate project evaluation is achieved.
Patent Information
- Application Number
- CN202411902052.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-12-23
- Publication Date
- 2025-05-30
AI Technical Summary
Project evaluation through manual review in the prior art leads to high labor and time costs, and the evaluation results are prone to deviations.
A comprehensive project evaluation method is adopted to obtain project categories and initial evaluation factors from previous evaluation reports, build evaluation networks and convolutional prediction neural network models, and automatically predict comprehensive evaluation of projects.
Intelligent and automated project evaluation has been realized, saving manpower and time costs, and improving the accuracy of evaluation.
Smart Images
Figure CN120069624A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of data processing, and particularly relates to a method and system for comprehensive project evaluation. Background Art
[0002] A project is a one-time independent task that organizes resources such as manpower, materials, and finances using various methods and arranges them according to relevant plans, with the aim of achieving goals defined by quantity and quality indicators.
[0003] Project evaluation refers to the assessment of the overall quality of a project, and its purpose is to screen out valuable projects from numerous projects.
[0004] The current method of project evaluation still uses manual evaluation. By manually reviewing the content of the project and then scoring the project, this manual review method not only requires a large amount of human and time costs, but also the active review criteria of the reviewers are difficult to unify, resulting in the evaluation results of the project quality being prone to deviation. Summary of the Invention
[0005] Aiming at the deficiencies of the prior art, the purpose of the present invention is to provide a method and system for comprehensive project evaluation, aiming to solve the technical problems in the prior art that project evaluation is achieved through manual review, which not only requires a large amount of human and time costs, but also the evaluation results are prone to deviation.
[0006] To achieve the above purpose, in the first aspect, the embodiments of the present application provide a method for comprehensive project evaluation, including the following steps:
[0007] Obtain a number of project categories and a number of initial evaluation factors from several previous evaluation reports, and construct several evaluation networks based on the project categories. The evaluation network includes a number of target evaluation factors;
[0008] Construct an initial convolutional prediction neural network model based on the random forest learning algorithm, obtain the component values corresponding to the target evaluation factors, and train the initial convolutional prediction neural network model through the component values to obtain the final prediction neural network model;
[0009] Obtain the category to be evaluated of the project to be evaluated, determine the target network from several of the evaluation networks based on the category to be evaluated, and select the target evaluation factors corresponding to the target network as the final evaluation factors;
[0010] Extract the target components corresponding to the final evaluation factors from the project to be evaluated, and obtain the comprehensive evaluation of the project to be evaluated through the target components and the final prediction neural network model.
[0011] Compared with the prior art, the beneficial effects of the present invention are as follows: By automatically identifying and extracting the target evaluation factors related to the evaluation results from the previous evaluation reports, a comprehensive and consistent evaluation standard system is constructed. By constructing the final prediction neural network model, it can learn from historical evaluation data and achieve automatic prediction of comprehensive evaluation, replacing manual evaluation with an intelligent and automated evaluation method, and obtaining the comprehensive evaluation with an objective evaluation standard. This not only saves labor costs and time costs, but also improves the accuracy of the comprehensive evaluation of the project to be evaluated.
[0012] Further, the step of obtaining a number of project categories and a number of initial evaluation factors from a number of previous evaluation reports includes:
[0013] Perform text cleaning on a number of the previous evaluation reports to filter tags and special characters, and then obtain a number of texts to be processed;
[0014] Perform word segmentation on the texts to be processed to obtain a number of words;
[0015] Perform semantic analysis on the previous evaluation reports through natural language processing technology and the BERT language model to select project categories and initial evaluation factors from a number of the words. Each of the previous evaluations corresponds to one project category and a number of the initial evaluation factors.
[0016] Furthermore, the step of constructing a number of evaluation networks based on the project categories, where the evaluation network includes a number of target evaluation factors, includes:
[0017] Perform duplicate removal processing on a number of the project categories to obtain a number of benchmark categories. Based on the benchmark categories, select a number of target evaluation factors from a number of the initial evaluation factors, and associate the number of target evaluation factors with the benchmark categories to form a number of initial networks;
[0018] Obtain factor values from the previous evaluation reports based on the target evaluation factors, and obtain the previous evaluation results of the previous evaluation reports. Determine the factor weights of the target evaluation factors based on the factor values and the previous evaluation results to form a number of evaluation networks.
[0019] Furthermore, the step of determining the factor weights of the target evaluation factors based on the factor values and the previous evaluation results includes:
[0020] Construct a graph neural network model, and collect evaluation standard values related to the target evaluation factors from a number of data sources through the graph neural network model;
[0021] Obtain a single evaluation result of the target evaluation factor through the factor value and the evaluation standard value;
[0022] Determine the factor weight of the target evaluation factor based on the single evaluation result and the previous evaluation result.
[0023] Furthermore, the single evaluation result includes a single positive evaluation and a single negative evaluation, the previous evaluation result includes a comprehensive positive evaluation and a comprehensive negative evaluation, and the step of determining the factor weight of the target evaluation factor based on the single evaluation result and the previous evaluation result includes:
[0024] Based on the number of previous evaluation reports corresponding to the reference category, obtain a first quantity corresponding to the single positive evaluation, a second quantity corresponding to the single negative evaluation, a third quantity corresponding to the comprehensive positive evaluation, and a fourth quantity corresponding to the comprehensive negative evaluation;
[0025] Determine a first weight through the first quantity and the third quantity, determine a second weight through the second quantity and the fourth quantity, and obtain the factor weight of the target evaluation factor through the first weight and the second weight.
[0026] Furthermore, the step of obtaining the component value corresponding to the target evaluation factor and training the initial convolutional prediction neural network model through the component value includes:
[0027] Based on the single evaluation result corresponding to the target evaluation factor, assign an initial value to the target evaluation factor, and obtain the component value through the initial value and the factor weight corresponding to the target evaluation factor;
[0028] Use the component value as the input value of the initial convolutional prediction neural network model, and use the previous evaluation result as the output value of the initial convolutional prediction neural network model to train the initial convolutional prediction neural network model.
[0029] Furthermore, the step of determining the target network from several evaluation networks based on the category to be evaluated includes:
[0030] Compare the category to be evaluated with the reference category to determine the target category from several reference categories;
[0031] Determine the evaluation network corresponding to the target category as the target network.
[0032] In a second aspect, an embodiment of the present application provides a project comprehensive evaluation system, which is applied to the project comprehensive evaluation method as described in the first aspect above. The system includes:
[0033] An analysis module for obtaining a number of project categories and a number of initial evaluation factors from several previous evaluation reports, constructing a number of evaluation networks based on the project categories, and the evaluation networks include a number of target evaluation factors;
[0034] A training module for constructing an initial convolutional prediction neural network model based on the random forest learning algorithm, obtaining component values corresponding to the target evaluation factors, and training the initial convolutional prediction neural network model through the component values to obtain a final prediction neural network model;
[0035] A selection module for obtaining the category to be evaluated of the project to be evaluated, determining a target network from a number of the evaluation networks based on the category to be evaluated, and selecting the target evaluation factors corresponding to the target network as the final evaluation factors;
[0036] An execution module for extracting target components corresponding to the final evaluation factors from the project to be evaluated, and obtaining a comprehensive evaluation of the project to be evaluated through the target components and the final prediction neural network model.
[0037] In a third aspect, an embodiment of the present application provides a computer, including a memory, a processor, and a computer program stored on the memory and executable on the processor. When the processor executes the computer program, the comprehensive project evaluation method described in the first aspect above is implemented.
[0038] In a fourth aspect, an embodiment of the present application provides a storage medium, on which a computer program is stored. When the computer program is executed by a processor, the comprehensive project evaluation method described in the first aspect above is implemented. BRIEF DESCRIPTION OF THE DRAWINGS
[0039] Figure 1 It is a flowchart of the comprehensive project evaluation method in the first embodiment of the present invention;
[0040] Figure 2 It is a structural block diagram of the comprehensive project evaluation system in the second embodiment of the present invention;
[0041] The following specific embodiments will further illustrate the present invention in conjunction with the above-mentioned drawings. DETAILED DESCRIPTION
[0042] For ease of understanding of the present invention, the present invention will be described more comprehensively below with reference to the relevant drawings. Several embodiments of the present invention are given in the drawings. However, the present invention can be implemented in many different forms and is not limited to the embodiments described herein. On the contrary, these embodiments are provided to make the disclosure of the present invention more thorough and comprehensive.
[0043] It should be noted that when an element is referred to as being "fixed to" another element, it can be directly on the other element or there can also be an intermediate element. When an element is considered to be "connected to" another element, it can be directly connected to the other element or there may be an intermediate element at the same time. The terms "vertical", "horizontal", "left", "right" and similar expressions used herein are for illustrative purposes only.
[0044] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those of ordinary skill in the technical field to which this invention belongs. The terms used in the description of the present invention herein are for the purpose of describing specific embodiments only and are not intended to limit the present invention. The term "and / or" used herein includes any and all combinations of one or more of the related listed items.
[0045] Please refer to Figure 1 , the project comprehensive evaluation method provided by the first embodiment of the present invention includes the following steps:
[0046] S10: Obtain a number of project categories and a number of initial evaluation factors from several previous evaluation reports, and construct a number of evaluation networks based on the project categories. The evaluation network includes a number of target evaluation factors;
[0047] The step S10 includes:
[0048] S110: Perform text cleaning on several of the previous evaluation reports to filter tags and special characters, and then obtain several texts to be processed;
[0049] S120: Perform word segmentation on the text to be processed to obtain several words;
[0050] S130: Perform semantic analysis on the previous evaluation reports through natural language processing technology and the BERT language model to select project categories and initial evaluation factors from several of the words. Each of the previous evaluations corresponds to one of the project categories and several of the initial evaluation factors;
[0051] By adopting the natural language processing technology and the BERT language model to deeply understand the semantic content in the previous evaluation reports, the recognition and extraction of the initial evaluation factors are completed accurately and automatically, ensuring the accuracy and applicability of the initial evaluation factors.
[0052] S140: Perform duplicate removal processing on several of the project categories to obtain several benchmark categories, select several target evaluation factors from several of the initial evaluation factors based on the benchmark categories, and associate several of the target evaluation factors with the benchmark categories to form several initial networks;
[0053] Understandably, for different previous evaluation reports, there will be some differences in the initial evaluation factors included therein, and selecting a number of the target evaluation factors can ensure a comprehensive selection of the influencing factors for the corresponding reference category.
[0054] S150: Obtain the factor values from the previous evaluation reports based on the target evaluation factors, and obtain the previous evaluation results of the previous evaluation reports. Determine the factor weights of the target evaluation factors based on the factor values and the previous evaluation results to form a number of evaluation networks;
[0055] Specifically, construct a graph neural network model, and collect the evaluation standard values related to the target evaluation factors from a number of data sources through the graph neural network model; obtain the single evaluation result of the target evaluation factor through the factor values and the evaluation standard values; determine the factor weight of the target evaluation factor through the single evaluation result and the previous evaluation result.
[0056] By constructing the graph neural network model, the information of different data sources can be identified and integrated in real time, and then the evaluation standard values can be updated at any time to ensure the accuracy of the evaluation. The single evaluation result includes a single positive evaluation and a single negative evaluation, and the previous evaluation result includes a comprehensive positive evaluation and a comprehensive negative evaluation. In this embodiment, the single positive evaluation is passing, the single negative evaluation is failing, the comprehensive positive evaluation is qualified, and the comprehensive negative evaluation is unqualified.
[0057] Understandably, the target evaluation factor can be a capital factor. When the factor value is 1 million and the evaluation standard value is 3 million, the single evaluation result is failing. When the factor value is 1 million and the evaluation standard value is 500,000, the single evaluation result is passing.
[0058] Furthermore, based on the number of the previous evaluation reports corresponding to the reference category, obtain the first quantity corresponding to the single positive evaluation, the second quantity corresponding to the single negative evaluation, the third quantity corresponding to the comprehensive positive evaluation, and the fourth quantity corresponding to the comprehensive negative evaluation; determine the first weight through the first quantity and the third quantity, and determine the second weight through the second quantity and the fourth quantity, and obtain the factor weight of the target evaluation factor through the first weight and the second weight.
[0059] If there are 10 of the previous evaluation reports, then for a certain target evaluation factor, there are 10 of the single evaluation results and 10 of the previous evaluation results. Further, by dividing the first quantity by the third quantity, a first weight is obtained, and by dividing the second quantity by the fourth quantity, a second weight is obtained. Then, the first weight and the second weight are averaged to obtain the factor weight.
[0060] S20: Based on the random forest learning algorithm, construct an initial convolutional prediction neural network model, obtain the component values corresponding to the target evaluation factor, and train the initial convolutional prediction neural network model with the component values to obtain the final prediction neural network model;
[0061] The step S20 includes:
[0062] S210: Based on the single evaluation results corresponding to the target evaluation factor, assign an initial value to the target evaluation factor, and obtain the component values through the initial value and the factor weight corresponding to the target evaluation factor;
[0063] The initial value is 0 or 1, and the product of the initial value and the factor weight is the component value.
[0064] S220: Use the component values as the input values of the initial convolutional prediction neural network model, and use the previous evaluation results as the output values of the initial convolutional prediction neural network model to train the initial convolutional prediction neural network model;
[0065] It can be understood that since there are several target evaluation factors, therefore, the component values corresponding to the number of target evaluation factors need to be combined into a data group, and then input into the initial convolutional prediction neural network model. The initial convolutional prediction neural network model outputs a predicted evaluation. A loss function is constructed through the predicted evaluation and the previous evaluation results, and the initial convolutional prediction neural network model is trained through the loss function. In this embodiment, the final prediction neural network model includes several sub-prediction models, and the sub-prediction models correspond to the reference categories one by one.
[0066] S30: Obtain the category to be evaluated of the project to be evaluated, determine the target network from several evaluation networks based on the category to be evaluated, and select the target evaluation factor corresponding to the target network as the final evaluation factor;
[0067] The step S30 includes:
[0068] S310: Compare the category to be evaluated with the reference categories to determine the target category from several reference categories;
[0069] S320: Determine the evaluation network corresponding to the target category as the target network.
[0070] Understandably, after determining the target network, the sub-prediction model corresponding to the target network can be correspondingly determined.
[0071] S40: Extract the target component corresponding to the final evaluation factor from the item to be evaluated, and obtain the comprehensive evaluation of the item to be evaluated through the target component and the final prediction neural network model;
[0072] The obtaining method of the target component is the same as that of the component value, and will not be elaborated here.
[0073] By automatically identifying and extracting the target evaluation factors related to the evaluation results from the previous evaluation reports, a comprehensive and consistent evaluation standard system is constructed. By constructing the final prediction neural network model, it is possible to learn from historical evaluation data and realize the automatic prediction of the comprehensive evaluation, replacing manual evaluation with an intelligent and automatic evaluation method, and obtaining the comprehensive evaluation with an objective evaluation standard. This not only saves labor costs and time costs, but also improves the accuracy of the comprehensive evaluation of the item to be evaluated.
[0074] Please refer to Figure 2 , the second embodiment of the present invention provides a project comprehensive evaluation system, which is applied to the project comprehensive evaluation method in the above embodiment, and those that have been described will not be elaborated again. As used below, terms such as "module", "unit", "sub-unit", etc. can be a combination of software and / or hardware that can achieve a predetermined function. Although the devices described in the following embodiments are preferably implemented in software, implementation in hardware, or a combination of software and hardware is also possible and contemplated.
[0075] The system includes:
[0076] An analysis module 10, configured to obtain a number of project categories and a number of initial evaluation factors from a number of previous evaluation reports, and construct a number of evaluation networks based on the project categories, where the evaluation network includes a number of target evaluation factors;
[0077] The analysis module 10 includes:
[0078] A first unit, configured to perform text cleaning on a number of the previous evaluation reports to filter tags and special characters, so as to obtain a number of texts to be processed;
[0079] A second unit, configured to perform word segmentation processing on the text to be processed to obtain a number of words;
[0080] The third unit is used to perform semantic analysis on the previous evaluation reports through natural language processing technology and the BERT language model to select item categories and initial evaluation factors from several of the words. Each of the previous evaluations corresponds to one of the item categories and several of the initial evaluation factors;
[0081] The fourth unit is used to perform duplicate removal processing on several of the item categories to obtain several benchmark categories, select several target evaluation factors from several of the initial evaluation factors based on the benchmark categories, and associate several of the target evaluation factors with the benchmark categories to form several initial networks;
[0082] The fifth unit is used to obtain factor values from the previous evaluation reports based on the target evaluation factors, and obtain the previous evaluation results of the previous evaluation reports. Based on the factor values and the previous evaluation results, determine the factor weights of the target evaluation factors to form several evaluation networks;
[0083] The fifth unit is specifically used to construct a graph neural network model, collect evaluation standard values related to the target evaluation factors from several data sources through the graph neural network model; obtain a single evaluation result of the target evaluation factor through the factor values and the evaluation standard values; determine the factor weight of the target evaluation factor through the single evaluation result and the previous evaluation result;
[0084] The fifth unit is further used to obtain the first quantity corresponding to the single positive evaluation, the second quantity corresponding to the single negative evaluation, the third quantity corresponding to the comprehensive positive evaluation, and the fourth quantity corresponding to the comprehensive negative evaluation based on the number of the previous evaluation reports corresponding to the benchmark categories; determine a first weight through the first quantity and the third quantity, determine a second weight through the second quantity and the fourth quantity, and obtain the factor weight of the target evaluation factor through the first weight and the second weight;
[0085] The training module 20 is used to construct an initial convolutional prediction neural network model based on the random forest learning algorithm, obtain component values corresponding to the target evaluation factors, and train the initial convolutional prediction neural network model through the component values to obtain a final prediction neural network model;
[0086] The training module 20 includes:
[0087] The sixth unit is used to assign an initial value to the target evaluation factor based on the single evaluation result corresponding to the target evaluation factor, and obtain component values through the initial value and the factor weight corresponding to the target evaluation factor;
[0088] A seventh unit is configured to use the component value as an input value of the initial convolutional prediction neural network model, and use the previous evaluation result as an output value of the initial convolutional prediction neural network model to train the initial convolutional prediction neural network model;
[0089] A selection module 30 is configured to obtain an evaluation category of an item to be evaluated, determine a target network from a plurality of the evaluation networks based on the evaluation category, and select the target evaluation factor corresponding to the target network as the final evaluation factor;
[0090] The selection module 30 includes:
[0091] An eighth unit is configured to compare the evaluation category with the reference categories to determine a target category from a plurality of the reference categories;
[0092] A ninth unit is configured to determine the evaluation network corresponding to the target category as the target network;
[0093] An execution module 40 is configured to extract a target component corresponding to the final evaluation factor from the item to be evaluated, and obtain a comprehensive evaluation of the item to be evaluated through the target component and the final prediction neural network model.
[0094] The present invention also provides a computer, including a memory, a processor, and a computer program stored on the memory and executable on the processor. When the processor executes the computer program, the item comprehensive evaluation method as described in the above technical solution is implemented.
[0095] The present invention also provides a storage medium, on which a computer program is stored. When the computer program is executed by a processor, the item comprehensive evaluation method as described in the above technical solution is implemented.
[0096] In the description of this specification, the descriptions with reference to the terms "one embodiment", "some embodiments", "example", "specific example", or "some examples", etc. mean that the specific features, structures, materials, or characteristics described in connection with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic representations of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials, or characteristics described can be combined in any one or more embodiments or examples in a suitable manner.
[0097] The above-described embodiments merely represent several implementation manners of the present invention. The description thereof is relatively specific and detailed, but it should not be construed as a limitation on the scope of the patent for the present invention. It should be noted that for those of ordinary skill in the art, without departing from the concept of the present invention, several modifications and improvements can still be made, and these all fall within the protection scope of the present invention. Therefore, the protection scope of the patent for the present invention shall be subject to the appended claims.
Claims
1. A comprehensive project evaluation method, characterized in that: The following steps are involved: Obtaining a number of project categories and a number of initial evaluation factors from a number of past evaluation reports, and constructing a number of evaluation networks based on the project categories, wherein the evaluation networks include a number of target evaluation factors; Constructing an initial convolutional prediction neural network model based on a random forest learning algorithm, obtaining component values corresponding to the target evaluation factor, and training the initial convolutional prediction neural network model through the component values to obtain a final prediction neural network model; Obtaining a category of the project to be evaluated, determining a target network from a plurality of evaluation networks based on the category, and selecting the target evaluation factor corresponding to the target network as a final evaluation factor; A target component corresponding to the final evaluation factor is extracted from the item to be evaluated, and a comprehensive evaluation of the item to be evaluated is obtained through the target component and the final prediction neural network model.
2. The project comprehensive evaluation method according to claim 1, characterized in that: The step of obtaining a number of project categories and a number of initial evaluation factors from a number of past evaluation reports includes: Performing text cleaning on some of the previous evaluation reports to filter out labels and special characters, thereby obtaining some texts to be processed; Performing word segmentation processing on the text to be processed to obtain a number of words; The previous evaluation reports are semantically analyzed by natural language processing technology and the BERT language model to select project categories and initial evaluation factors from the several words, and each of the previous evaluations includes a corresponding project category and several initial evaluation factors.
3. The project comprehensive evaluation method according to claim 2, characterized in that: The step of constructing a plurality of evaluation networks based on the project categories, wherein the evaluation networks include a plurality of target evaluation factors, comprises: Deduplication processing is performed on a plurality of the project categories to obtain a plurality of benchmark categories, a plurality of target evaluation factors are selected from a plurality of the initial evaluation factors based on the benchmark categories, and the plurality of the target evaluation factors are associated with the benchmark categories to form a plurality of initial networks; Based on the target evaluation factor, factor values are obtained from the previous evaluation reports, and previous evaluation results of the previous evaluation reports are obtained. Based on the factor values and the previous evaluation results, factor weights of the target evaluation factors are determined to form a plurality of evaluation networks.
4. The project comprehensive evaluation method according to claim 3 is characterized in that: The step of determining the factor weight of the target evaluation factor based on the factor value and the previous evaluation results includes: Constructing a graph neural network model, and collecting evaluation standard values related to the target evaluation factor from a plurality of data sources through the graph neural network model; Obtaining a single evaluation result of the target evaluation factor through the factor value and the evaluation standard value; The factor weight of the target evaluation factor is determined by the single evaluation result and the previous evaluation results.
5. The project comprehensive evaluation method according to claim 4 is characterized in that: The single evaluation result includes a single positive evaluation and a single negative evaluation, the past evaluation results include a comprehensive positive evaluation and a comprehensive negative evaluation, and the step of determining the factor weight of the target evaluation factor by using the single evaluation result and the past evaluation results includes: Based on the number of the past evaluation reports corresponding to the benchmark category, obtaining the first number corresponding to the single positive evaluation, the second number corresponding to the single negative evaluation, the third number corresponding to the comprehensive positive evaluation, and the fourth number corresponding to the comprehensive negative evaluation; A first weight is determined by the first quantity and the third quantity, and a second weight is determined by the second quantity and the fourth quantity, and a factor weight of the target evaluation factor is obtained by the first weight and the second weight.
6. The project comprehensive evaluation method according to claim 4, characterized in that: The step of obtaining component values corresponding to the target evaluation factors and training the initial convolutional prediction neural network model by using the component values comprises: Based on the single evaluation result corresponding to the target evaluation factor, assigning an initial value to the target evaluation factor, and obtaining a component value through the initial value and the factor weight corresponding to the target evaluation factor; The component values are used as input values of the initial convolutional prediction neural network model, and the previous evaluation results are used as output values of the initial convolutional prediction neural network model to train the initial convolutional prediction neural network model.
7. The project comprehensive evaluation method according to claim 1, characterized in that: The step of determining a target network from a plurality of evaluation networks based on the category to be evaluated comprises: Comparing the category to be evaluated with the benchmark category to determine a target category from a plurality of the benchmark categories; The evaluation network corresponding to the target category is determined as a target network.
8. A project comprehensive evaluation system, applied to the project comprehensive evaluation method according to any one of claims 1 to 7, characterized in that: The system comprises: An analysis module, used to obtain a number of project categories and a number of initial evaluation factors from a number of past evaluation reports, and to construct a number of evaluation networks based on the project categories, wherein the evaluation networks include a number of target evaluation factors; A training module, used to construct an initial convolutional prediction neural network model based on a random forest learning algorithm, obtain component values corresponding to the target evaluation factors, and train the initial convolutional prediction neural network model through the component values to obtain a final prediction neural network model; A selection module, configured to obtain a category of the project to be evaluated, determine a target network from a plurality of evaluation networks based on the category, and select the target evaluation factor corresponding to the target network as a final evaluation factor; An execution module is used to extract a target component corresponding to the final evaluation factor from the project to be evaluated, and obtain a comprehensive evaluation of the project to be evaluated through the target component and the final prediction neural network model.
9. A computer comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the project comprehensive evaluation method according to any one of claims 1 to 7 is implemented.
10. A storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the project comprehensive evaluation method according to any one of claims 1 to 7 is implemented.