Sentimental Analysis using Conceptual Clustering Approach

• Main Authors: Yun-PengSu, 蘇雲鵬
• Other Authors: Sheng-Tun Li
• Format: Others
• Language: zh-TW
• Published: 2012
• Online Access: http://ndltd.ncl.edu.tw/handle/60840040644759787494

碩士 === 國立成功大學 === 工業與資訊管理學系專班 === 100 === Human beings are social creatures. Thus, getting along with others must have got some interactions. In other words, one is able to clearly comprehend the information that the other would like to deliver to by human communication. In these delivered information, some messages such as identification, causes of occurrences and sentiments at that time would be hidden except literal meanings. For the sake of the immediateness and convenience, questionnaires which are frequently used in the project analyzed region must occasionally use a few and limited words to express and to deliver the matters and statuses at that moment. In questionnaires, the sentences or the articles are almost brief. Therefore, we may miss plenty of important information frequently if we just emphasize the expression of literal meanings. Enormous amount of data and consumption of human resources may cause inefficiency if we analyze and recognize by manpower. This trouble is solved because of the rapid development of computer software and hardware. Computers has substituted for the usage of manpower In the research, we probe into fuzzy formal context and membership value between objects and attributes in fuzzy formal concept. We expect to build up a bridge between fuzzy formal context and concept lattice with an automatic mode. So that we can discover the hidden sentiments in documents and categorize them. In order to obtain a better parameter combination, we proceed an experimental design by choosing Taguchi Method according to each parameter value in the research. Nevertheless, we hope that traditional manpower categorization can be replaced by this research someday in the future.