CN-122022704-A - Auxiliary evaluation mark and intelligent report generation system based on large model

CN122022704ACN 122022704 ACN122022704 ACN 122022704ACN-122022704-A

Abstract

The invention relates to the technical field of bid evaluation, and discloses an auxiliary bid evaluation and intelligent report generation system based on a large model, wherein a model training module acquires multidimensional bid evaluation indexes of a plurality of bid files, inputs the multidimensional bid evaluation indexes into a pre-trained large language evaluation model and outputs multidimensional evaluation values of each multidimensional bid evaluation index; the method comprises the steps of combining the multi-dimensional evaluation values of the same type corresponding to each bidding document to obtain a plurality of multi-dimensional evaluation value groups, calculating document evaluation measurement values of the plurality of bidding documents, analyzing all the document evaluation measurement values by a comprehensive evaluation module, calculating comprehensive document evaluation coefficients of the bidding documents based on analysis results, generating bid evaluation reports corresponding to all the bidding documents by a report generation module based on the comprehensive document evaluation coefficients, and carrying out cooperative bid evaluation on a plurality of associated bidding documents so as to solve the problems of low efficiency and insufficient accuracy in the conventional bid evaluation method when combined bidding is processed, and improving the bid evaluation efficiency and accuracy.

Inventors

TIAN SHIDONG
CHEN GUOXUAN
JING YANXIN

Assignees

华能招采数字科技有限公司

Dates

Publication Date: 20260512
Application Date: 20251201

Claims (10)

1. Auxiliary comment and intelligent report generation system based on big model, characterized by comprising: the model training module is used for acquiring multi-dimensional bid evaluation indexes of a plurality of bidding documents, inputting the multi-dimensional bid evaluation indexes into a pre-trained large language evaluation model, and outputting multi-dimensional evaluation values corresponding to each multi-dimensional bid evaluation index; The index evaluation module is used for combining the multi-dimensional evaluation values of the same type corresponding to each bidding document to obtain a plurality of multi-dimensional evaluation value groups, and calculating document evaluation metric values of a plurality of bidding documents according to the multi-dimensional evaluation value groups; The comprehensive evaluation module is used for analyzing all the file evaluation metric values and calculating the comprehensive file evaluation coefficients of the bidding files based on the analysis result; and the report generation module is used for generating bid evaluation reports corresponding to all bidding documents based on the comprehensive document evaluation coefficients.
2. The large model based auxiliary comment and intelligence report generating system of claim 1, further comprising: And the index processing module is used for carrying out index processing on all the multi-dimensional index, wherein the index processing comprises the steps of deleting repeated multi-dimensional index, deleting wrong multi-dimensional index and deleting invalid multi-dimensional index.
3. The large model-based auxiliary bid evaluation and intelligent report generation system of claim 1, wherein the model training module is configured to: The model training module is used for acquiring a bidding document and a bid evaluation standard document; the model training module is used for constructing instruction-answer pairs based on the bidding documents and the bid evaluation standard documents and generating a domain-specific training data set; The model training module is used for training the general large language evaluation model by utilizing the field specific training data set to obtain the large language evaluation model.
4. The large model-based auxiliary bid evaluation and intelligent report generation system of claim 1, wherein the index evaluation module is configured to: The index evaluation module is used for sorting all the multidimensional evaluation values based on the acquisition time sequence; The index evaluation module is used for extracting a first multi-dimensional evaluation value and a tail multi-dimensional evaluation value and calculating the overall change factor of the multi-dimensional evaluation value group according to the first multi-dimensional evaluation value and the tail multi-dimensional evaluation value; the index evaluation module is used for calculating the progressive variation factor of the multi-dimensional evaluation value group based on the residual multi-dimensional evaluation values; And the index evaluation module is used for weighting and summing the integral change factor and the progressive change factor to obtain the file evaluation metric value of the bidding file.
5. The large model based auxiliary bid evaluation and intelligent report generation system of claim 4, wherein the index evaluation module is configured to: The index evaluation module is used for calculating the absolute value of the difference value of the first multi-dimensional evaluation value and the last multi-dimensional evaluation value to be used as the difference of the multi-dimensional evaluation values; The index evaluation module is used for determining the average value of all the multi-dimensional evaluation values, and taking the ratio of the difference of the multi-dimensional evaluation values to the average value as the overall variation factor of the multi-dimensional evaluation value group.
6. The large model based auxiliary bid evaluation and intelligent report generation system of claim 4, wherein the index evaluation module is configured to: the index evaluation module is used for randomly extracting a multi-dimensional evaluation value from the rest multi-dimensional evaluation values to serve as a standard multi-dimensional evaluation value; the index evaluation module is used for presetting a first value and a second value; The index evaluation module is used for acquiring a left neighborhood multidimensional evaluation value corresponding to the standard multidimensional evaluation value based on the first value; the index evaluation module is used for acquiring a right neighborhood multi-dimensional evaluation value corresponding to the standard multi-dimensional evaluation value based on the second value; The index evaluation module is used for combining the left neighborhood multi-dimensional evaluation value, the right neighborhood multi-dimensional evaluation value and the standard multi-dimensional evaluation value to obtain a multi-dimensional evaluation value set; The index evaluation module is used for calculating a neighborhood fluctuation value of the standard multi-dimensional evaluation value based on the multi-dimensional evaluation value set; The index evaluation module is used for calculating neighborhood fluctuation values of each residual multi-dimensional evaluation value; The index evaluation module is used for performing curve fitting on all the neighborhood fluctuation values to obtain a neighborhood fluctuation value curve, and extracting the maximum curve slope; The index evaluation module is used for calculating standard deviation of all neighborhood fluctuation values and taking the product value of the standard deviation and the maximum curve slope as a progressive variation factor of the multi-dimensional evaluation value group.
7. The large model based auxiliary comment and intelligence report generating system of claim 6 wherein said index evaluation module is configured to: the index evaluation module is used for calculating a neighborhood fluctuation value of the standard multidimensional evaluation value according to the following formula: ; Wherein q is the neighborhood fluctuation value of the standard multi-dimensional evaluation value, w is the number of the multi-dimensional evaluation values in the multi-dimensional evaluation value set, r w is the w-th multi-dimensional evaluation value in the multi-dimensional evaluation value set, r min is the minimum multi-dimensional evaluation value in the multi-dimensional evaluation value set, and r max is the maximum multi-dimensional evaluation value in the multi-dimensional evaluation value set.
8. The large model based auxiliary evaluation and intelligent report generating system according to claim 1, wherein the comprehensive evaluation module is configured to: the comprehensive evaluation module is used for determining the average value corresponding to all the file evaluation measurement values and taking the average value as the characteristic file evaluation measurement value; The comprehensive evaluation module is used for extracting file evaluation metric values smaller than the characteristic file evaluation metric values and counting the corresponding first file evaluation metric value quantity; The comprehensive evaluation module is used for extracting file evaluation metric values which are equal to the characteristic file evaluation metric values and counting the number of corresponding second file evaluation metric values; the comprehensive evaluation module is used for extracting file evaluation metric values larger than the characteristic file evaluation metric values and counting the number of corresponding third file evaluation metric values; The comprehensive evaluation module is used for calculating the comprehensive file evaluation coefficient of the bidding file according to the first file evaluation metric value quantity, the second file evaluation metric value quantity and the third file evaluation metric value quantity.
9. The large model based auxiliary bid evaluation and intelligent report generation system of claim 8, wherein the comprehensive evaluation module is configured to: the comprehensive evaluation module is used for calculating the comprehensive file evaluation coefficient of the bidding file according to the following formula: ; wherein v is the comprehensive file evaluation coefficient of the bidding file, j1 is the first file evaluation metric value, j2 is the second file evaluation metric value, and j3 is the third file evaluation metric value.
10. The large model based auxiliary comment and intelligence report generating system of claim 1 wherein said report generating module is configured to: The report generation module is used for automatically generating a bid evaluation report with complete structure based on the comprehensive file evaluation coefficient, the bid key information and a preset report template.

Description

Auxiliary evaluation mark and intelligent report generation system based on large model Technical Field The invention relates to the technical field of bid evaluation, in particular to an auxiliary bid evaluation and intelligent report generation system based on a large model. Background In the fields of engineering construction projects, service bidding and the like, evaluation is a core link for ensuring project quality, controlling cost and selecting an optimal partner. Most of the existing bid evaluation methods are independently carried out based on single bid segments. The bid evaluation committee performs independent scoring or evaluation on each bidding document according to preset bid evaluation standards (such as price, technical scheme, enterprise qualification, performance and the like), and finally determines the bid-winning candidate of the bidding segment. Although the flow of the method is clear, the method has obvious defects when processing a plurality of associated segments: when bidders bid in a "combined" way (i.e., bid on multiple bid sections simultaneously and offer combined offers), existing systems lack efficient mathematical models and calculation engines to handle such complex, nonlinear scoring relationships, and it is difficult to perform efficient and accurate comprehensive calculations while ensuring fairness. Disclosure of Invention The embodiment of the invention provides an auxiliary bid evaluation and intelligent report generation system based on a large model, which can carry out cooperative bid evaluation on a plurality of associated bid files so as to solve the problems of low efficiency and insufficient accuracy in the conventional bid evaluation method when combined bids are processed, and improve the bid evaluation efficiency and accuracy. In order to achieve the above object, the present invention provides a large model-based auxiliary comment and intelligent report generation system, comprising: the model training module is used for acquiring multi-dimensional bid evaluation indexes of a plurality of bidding documents, inputting the multi-dimensional bid evaluation indexes into a pre-trained large language evaluation model, and outputting multi-dimensional evaluation values corresponding to each multi-dimensional bid evaluation index; The index evaluation module is used for combining the multi-dimensional evaluation values of the same type corresponding to each bidding document to obtain a plurality of multi-dimensional evaluation value groups, and calculating document evaluation metric values of a plurality of bidding documents according to the multi-dimensional evaluation value groups; The comprehensive evaluation module is used for analyzing all the file evaluation metric values and calculating the comprehensive file evaluation coefficients of the bidding files based on the analysis result; and the report generation module is used for generating bid evaluation reports corresponding to all bidding documents based on the comprehensive document evaluation coefficients. Further, the method further comprises the following steps: And the index processing module is used for carrying out index processing on all the multi-dimensional index, wherein the index processing comprises the steps of deleting repeated multi-dimensional index, deleting wrong multi-dimensional index and deleting invalid multi-dimensional index. Further, the model training module is configured to: The model training module is used for acquiring a bidding document and a bid evaluation standard document; the model training module is used for constructing instruction-answer pairs based on the bidding documents and the bid evaluation standard documents and generating a domain-specific training data set; The model training module is used for training the general large language evaluation model by utilizing the field specific training data set to obtain the large language evaluation model. Further, the index evaluation module is used for: The index evaluation module is used for sorting all the multidimensional evaluation values based on the acquisition time sequence; The index evaluation module is used for extracting a first multi-dimensional evaluation value and a tail multi-dimensional evaluation value and calculating the overall change factor of the multi-dimensional evaluation value group according to the first multi-dimensional evaluation value and the tail multi-dimensional evaluation value; the index evaluation module is used for calculating the progressive variation factor of the multi-dimensional evaluation value group based on the residual multi-dimensional evaluation values; And the index evaluation module is used for weighting and summing the integral change factor and the progressive change factor to obtain the file evaluation metric value of the bidding file. Further, the index evaluation module is used for: The index evaluation module is used for calculating the absolute value of the diffe