電子化政府查報案件自動分類設計之研究-以台北市政府為例

A Study of Automatic Classification and Design of E-Government Case Investigation – Taking Taipei City Government as an Example

蕭群殷(Chun-Yin Hsiao)

指導教授 ： 蕭瑞祥

淡江大學/商管學院/資訊管理學系碩士在職專班/碩士(2017年)

https://doi.org/10.6846/TKU.2017.00727

摘要

『電子化政府』是藉由資訊科技的應用以提升政府行政效率與服務品質。近年來每年市容通報案件逐年成長到年近20萬件的數量，迫使政府需投入大量的人力來處理，使得市政效率下降與成本加劇，因此建立高效率、操作容易與人性化界面的網站平台，讓市民隨時通報市容改善案件，提升市府改善市容工程的效率與品質已是重要的關鍵課題。
本研究以文件自動分類切入，透過系統雛形法來做為研究方法，使用大量的市容查報案件的資訊進行分類模型訓練，並輔以使用者體驗來進行結果探討，實驗結果獲得87.04%的準確率高於傳統的分類60%。期望透過本研究可以提供市容查報系統業者進行系統開發時的技術參考。

關鍵字

機器學習 ； 文件自動分類 ； 市容查報 ； 系統雛形

並列摘要

The "e-government" aims at enhancing the efficiency of government administration and service quality through the application of information technology. In recent years, the reported cases about the city’s appearance has been gradually increasing year by year to a number of nearly 200,000 for a year, forcing the government to invest a lot of manpower to deal with it. This has thus caused a decline in the efficiency of municipal administration and increase in relevant costs. Therefore, in terms of promoting the efficiency of handling reports and improvement of the city’s appearance, it is an important and critical task for the government to establish an efficient, easy-to-operate website platform which can be used by residents to report related issues anytime.

In this study, system prototyping was adopted as the research methodology based on automatic text categorization of the large quantity of reports about the appearance of the city for conducting classification model training. Moreover, it also explored the results according to user experience. The results of the experiment showed that it acquired 87.04% accuracy rate, which was higher than the accuracy rate by conventional classification which is at 60%. It is expected that the study could provide technical reference for the system development service providers.

並列關鍵字

Machine learning ； Document classification ； City's checking and feedback reports ； System prototyping

參考文獻

[8] 陳祥、許嘉文，2003，「各國電子化政府整合型入口網站功能比較分析」，研考雙月刊，第27卷，第5期，頁102-115。

[7] 陳世榮，2015，「社會科學研究中的文字探勘應用：以文意為基礎的文件分類及其問題」，人文及社會科學集刊，第二十七卷第四期，頁683-718。

[14] 韓歆儀，2004，「應用兩階段分類法提昇SVM法之分類準確率」，國立成功大學工業與資訊管理研究所碩士論文。

[16] Ghazinoory, S., Aliahmadi A., Namdarzangeneh, S., & Ghodsypour, S. H. 2007. "Using AHP and L.P. for choosing the best alternatives based the gap analysis," Applied Mathematics and Computation vol. 184, pp.316–321.

[21] Joachims, T. 1998. "Text Categorization with Support Vector Machines: Learning with Many Relevant Features," Proceedings of {ECML}-98, 10th European Conference on Machine Learning, pp.137-142.