Monitoring system of opinion information based on web 2.0

초록

The suggesting system automatically extracts and analyses user opinion information from web content that is scattered on various websites of the Internet thereby provides an opinion search service that enables checking statistics and search by positive and negative opinions. As a result, opinion search users may use a search and monitoring system easily at a glance of other users' opinions on specific key words; it also offers a function that extracts and analyses opinions from content. In order words, the system can dramatically reduce the time and costs of individuals or groups in searching for opinions of other users. The composition of the paper suggests issues on related researches and described system confirmation as well as methods for opinion information automatic extraction suggested in the paper. The full details used for the methods of opinion information automatic extraction and described in the paper are data collection processing, opinion/non-opinion automatic construction, opinion expression machine learning, indexing transaction, and opinion search processing. Comparison experiments were implemented for suggested techniques and other technologies whereby proved excellence of actual performance. The study also implemented the performance evaluation of a function that extracts positive and negative opinion information. We tested movie review scripts as a typical application case as well as Wikipedia's test data and performed a comparative analysis of the result. ©2014 International Information Institute.

키워드