基于RASA的智能气象问答服务技术的设计研究  

Design and Research of Intelligent Weather Question-Answering Service Technology Based on RASA

作　　者：杨涵 徐虹[1] YANG Han;XU Hong(Chengdu University of Infomation Technology,Chengdu,Sichuan 610225)

机构地区：[1]成都信息工程大学,四川成都610225

出　　处：《长江信息通信》2024年第4期12-16,共5页Changjiang Information & Communications

摘　　要：本研究基于Rasa开源问答系统框架和Flask轻量级Web服务框架,设计面向任务的智能气象问答系统。文章重点介绍了Rasa框架的关键模块RasaNLU,并结合自然语言理解中的意图和命名实体,设计了对话追踪和对话策略优化,相较于Rasa框架自带的RasaCore,在响应速度上得到了提升,同时详细描述了智能问答在气象可视化上的应用。This study focuses on designing a task-oriented intelligent weather question-answering system based on the Rasa open-source question-answering system framework and the Flask lightweight web service framework.The paper cmphasizes the key module RasaNLU within the Rasa framework,incorporating natural language understanding with intent and named cntity rccognition for dialoguc tracking and optimization of conversation strategy,Compared to the built-in RasaCore,improvements in response speed have been achieved,The study also provides a detailed description of the application of intelligent question-answering in weather visualization.

关 键 词：问答系统 RASA 自然语言理解 对话管理 气象可视化 

分 类 号：TP311[自动化与计算机技术—计算机软件与理论]