Quotation and Assessment System

Abstract

We provide an estimation and assessment system that can be widely used for various types of estimates and that can present the basis and explanations for the assessment results. [Solution] The system includes an estimate conversion unit 10 that converts the contents of the estimate document 11 into text and records it as structured data 12; an estimate evaluation unit 20 that predicts an appropriate price from the contents of the structured data 12 based on an estimate model 42 generated by machine learning based on past estimate data 41 and outputs it as an estimate evaluation result 21; and a report creation unit 30 that generates explanatory text related to the estimate evaluation result 21 using LLM2 based on the estimate document 11 and the estimate evaluation result 21, and outputs it as an estimate evaluation report 31 that includes the contents combined with the estimate evaluation result 21.

Claims

1. A quotation evaluation system for assessing the content of quotations in quotation documents, An estimate conversion unit that converts the contents of the aforementioned estimate document into text, converts it into structured data, and records it, An estimation evaluation unit predicts an appropriate price from the content of the structured data and outputs it as an estimation evaluation result based on an estimation model generated by machine learning based on past estimation data, A cost assessment system comprising: a report generation unit that generates explanatory text related to the cost assessment results using a large-scale language model (hereinafter referred to as "LLM") based on the cost assessment materials and the cost assessment results, and outputs it as a cost assessment report that includes the content combined with the cost assessment results.

2. In the estimate assessment system described in claim 1, The estimate conversion unit is an estimate assessment system that, when recording the structured data, records the estimate items in the structured data after matching them to the names of existing estimate items.

3. In the estimate assessment system described in claim 1, The estimation evaluation unit determines and visualizes the contribution of each estimation item in the structured data to the appropriate price using a predetermined method related to interpretable machine learning, and outputs this as part of the estimation evaluation results, thereby providing an estimation assessment system.

4. In the estimate assessment system described in claim 1, The aforementioned report creation unit is an estimate evaluation system that extracts and outputs past estimate documents similar in content to the aforementioned estimate document.

5. In the estimate assessment system described in claim 1, The aforementioned estimation model has multiple variations depending on the number of features, The estimation evaluation unit selects the estimation model in which the degree of match between the estimation items and the features in the structured data is greater than a predetermined value and the number of features is the largest, and predicts the appropriate price based on the selected estimation model.

Images