联系我们

论文查询结果	返回搜索

论文编号：	13066
作者编号：	2120193009
上传时间：	2022/6/5 10:46:18
中文题目：	政府网站智能问答系统服务质量评价指标体系构建研究
英文题目：	Research on the Evaluation Index System of Service Quality for Intelligent Question Answering System on Government Portals
指导老师：	王芳
中文关键字：	政府网站；智能问答系统；评价体系；测试题集；服务质量
英文关键字：	Government Portal; Intelligent Question Answering System; Evaluation System; Test Question Collection; Quality of Service
中文摘要：	随着人工智能与自然语言处理技术的发展，智能问答技术有了较大突破，该技术逐渐被应用于越来越多的领域。为打造智慧政府新形态，政府网站也陆续为用户提供智能问答服务，然而诸多网站中的问答系统表现良莠不齐，现有研究又缺乏对其评价的合适标准，探讨一套适用于政府网站智能问答系统的评价指标体系，有利于提升政府信息服务水平，促进政府服务的智慧化转型。 基于此，本文以新公共服务理论为导向，首先使用文献研究法对智能问答系统、政府网站信息化服务评价、搜索引擎评价等相关文献进行回顾，利用案例分析法对上海、浙江等四地具有特色的问答系统进行了研究，提取了可供政府网站智能问答系统使用的评测指标。而后，本文通过个人分析以及德尔菲法确立了评价体系以及指标权重，最终创建的政府网站智能问答评价体系包含问题解决质量、服务交互质量、基础建设质量 3 个一级指标以及高频问题解答率、同义问题解答率、交互次数等 18 个细化指标。与此同时，本文将政府网站中的公民留言进行文本处理，通过与国务院颁布的高频政务事项对比，编制了 60 道标准化的高频测试问题。随后，根据汉语表达方式、用户输入习惯等特点对高 频问题进行改编，最终形成了与评价体系配套的问题测试合集。最后，本文利用构建的评价指标体系与配套测试问题集对 30 个省级政府网站中的智能问答系统的服务质量进行了实证评测研究。 实证研究证明了本文评价体系的合理性与可操作性，同时测试结果也表明，当前仅有 30%的政府网站问答系统得分处于及格线以上，而大部分系统的表现都不尽人意。针对此现象，本文从系统解答能力、交互能力以及基础建设三个方面分别提出改进措施，以期提升智能问答服务水平，增强公民对于政府服务智慧化转型的信心。图 30 幅，表 23 个，参考文献 101 篇。
英文摘要：	With the development of artificial intelligence and natural language processing, there has been a breakthrough in intelligent question answering technology, which is gradually applied to more and more fields. To create a new form of smart government, government portals have been providing users with intelligent question answering services. However, the performance of question answering systems in many portals is mixed, and existing research lacks suitable criteria for evaluation. Therefore,exploring an evaluation index system for intelligent question answering systems on government portals is beneficial to improving the level of government information services and promoting the transformation of government services’ becoming more intelligent. On this basis, the paper takes the new public service theory as the guide and firstly uses the literature research method to review the literature related to intelligent question answering systems, service evaluation of government portals informatization,and search engines evaluation. Next, this thesis studies the question answering systems with characteristics in four places such as Shanghai and Zhejiang by case analysis method and extracts the evaluation indexes that can be used for intelligent question answering systems on government portals. Then, the evaluation system and index weight are established through personal analysis and the Delphi method. This evaluation system for question answering systems on government portals includes 3 first-grade indexes, such as the quality of problem solving, quality of service interaction, and quality of infrastructure, and 18 detailed indexes, such as the answer rate of high frequency questions, answer rate of synonymous questions and beauty degree of the interface. Meanwhile, this paper processes the text of citizens' messages left on government portals and compiles 60 high-frequency standard test questions by comparing those messages with the high-frequency government affairs issued by the State Council. Subsequently, the high frequency questions are adapted according to the Chinese expressions, users’ input habits, and other characteristics, and formed a test question collection that matches the evaluation system. Finally, this paper conducts an empirical evaluation study on the service quality of the intelligent question answering system on 30 provincial government portals using the evaluation index system and the collection of test questions. The empirical study proves the rationality and operability of the evaluation system established in this paper, while the test results also show that only 30% of the current government portals’ intelligent question answering systems score above the passing line, and most of the rest perform unsatisfactorily. In response to the phenomenon, this paper proposes improvement measures in three aspects: system answer capability, interaction capability and infrastructure construction, aiming to promote the level of intelligent question answering services and enhance citizens'confidence in the transformation of government services’ becoming more intelligent.This paper includes 30 Figures, 23 Tables, 101 References.
查看全文：	预览  下载(下载需要进行登录)