A Study on Text Classification for Mass Rapid Transit News Articles

• Main Authors: Wen-Lan Li, 李文嵐
• Other Authors: Liang-Chih Yu
• Format: Others
• Language: zh-TW
• Published: 2019
• Online Access: http://ndltd.ncl.edu.tw/handle/4gxfzn

碩士 === 元智大學 === 資訊管理學系 === 107 === With the development of economy, the advancement of science and technology and the change of life style, the emergence of the MRT system brings people fast and convenient traffic, so it attracts more and more people to take the MRT. However, in recent years, consumers' awareness has risen, customers value their own rights and interests, and establishing a good relationship with customers is one of the keys to long-term success. The MRT has a passenger complaint channel and has a coordinator to handle the appeal case, but there may be many complicated cases every day. The case must be classified by manual reading and then handed over to a coordinator. Work is time consuming and labor intensive. Therefore, this study collects MRT related news articles, and through pre-processing and manual tagging. It uses word embedding to convert text into vectors and constructs models: Support Vector Machine, Random Forest, Long Short-Term Memory, and Convolutional Neural Networks to classify MRT news articles. Last, evaluates the effectiveness of the model and provides the issues which can refer in the future. In advance, the research implements an automatic text classification system.