基于大语言模型的竞品车型配置问答系统设计与应用研究
Design and Application Research of Competitive Vehicle Configuration Question Answering System Based on Big Language Model
雷天凤 1张永 2龚春忠 1周伟明1
作者信息
- 1. 合众新能源汽车股份有限公司,嘉兴 314000
- 2. 合众新能源汽车股份有限公司,嘉兴 314000;浙江工业大学,杭州 310000
- 折叠
摘要
本研究致力于优化汽车研发工程师对大量车型配置信息的研究和分析过程.通过设计一款基于大语言模型的智能问答系统,工程师能够以自然语言形式获取车型配置信息的统计和可视化结果,避免了逐步筛选或使用数据分析工具的繁琐步骤,提高了竞品数据分析的效率.考虑到目前大语言模型在汽车预研工作中的限制,我们采用案例分析法,以已构建的车型配置表为基础,制定了六步骤的研究过程,包括系统设计、大语言模型选择、数据库构建、自然语言解析、SQL执行与可视化,以及最终结果的封装.成功实现了基于大语言模型的智能问答系统,为汽车工程师提供了一个便捷的竞品问题解答平台,填补了在汽车领域对车型配置信息利用方面的不足,使工程师能够更迅速地获取他们所需的关键问题答案.
Abstract
This study aims to optimize the research and analysis process of automotive R&D engineers on a large amount of vehicle configuration information.By designing an intelligent question answering system based on a big language model,engineers can obtain statistical and visual results of vehicle configuration information in natural language form,avoiding the tedious steps of gradually screening or using data analysis tools,and improving the efficiency of competitor data analysis.Considering the limitations of the current big language model in automotive pre research work,we adopted a case study method and developed a six step research process based on the constructed vehicle configuration table,including system design,big language model selection,database construction,natural language parsing,SQL execution and visualization,and encapsulation of the final results.Successfully implemented an intelligent question answering system based on a large language model,providing a convenient platform for automotive engineers to answer competitive questions,filling the gap in the utilization of vehicle configuration information in the automotive field,and enabling engineers to quickly obtain the key question answers they need.
关键词
问答系统/大语言模型/车型配置信息/竞品分析Key words
Q&A System/Big Language Model/Vehicle Configuration Information/Competitive Analysis引用本文复制引用
出版年
2024
NETL
NSTL
万方数据