| Summary: | 碩士 === 華梵大學 === 資訊管理學系碩士班 === 102 === With the increasing popularity of Internet, information dependence on the Internet also gives rise. How to use the electronic media to spread information quickly and efficiently obtain useful information on their needs, is an important issue.
This study discussed the classification using the news headlines. The automatic document classification (text classification) method was adopted. The word processing and the frequency information of words were used to extract the attributes (features) for the classification of the news articles using the Weka data mining system.
We discussed the impacts of the ratio of the files (1:1, 1:5, 1:10) and the numbers of attributes (128, 256, 512, 1024, 2048) for the classification of news. The classifications using the contents and the headlines of the news articles were also compared. The experimental results show that the classification of the news content gets the accuracy of 96.3636%, and slightly higher than the correct rate of 93.6364% for the headlines. The difference is 2.0454%. The results show that the proposed method can be used for the classification of the real-time electronic news using the title field of the news.
Keywords: text classification, news headlines, data mining
|
|---|
