期刊名称:International Journal of Computer and Information Technology
印刷版ISSN:2279-0764
出版年度:2015
卷号:4
期号:2
出版社:International Journal of Computer and Information Technology
摘要:In modern era Text Mining and Sentiment investigation is a growing research area, covering over several disciplines such as data mining, text mining, etc. Text mining is a vital art of extracting the texts from the huge set of text set or reviews. Sentiment investigation is a category of natural language processing for tracking the mood of the public about a specific product or theme. The present works of text mining used Sentiwordnet as a lexical resource. The major drawback of this present Sentiwordnet is non-determination of total count, such as it doesn't offer the specific number of +1, -1 and 0 total words. These data and facts are necessary because without these details of total count if further data mining techniques are applied, it may give erroneous results. To facilitate the text mining task, this work focus on design of Developed Sentiwordnet so that it can produce the count of total words by distinguishing them into +1, - 1 and 0 words. Experiments are conducted on standard movie review and product review datasets. These works also make use of Stanford POS tagger for labeling the dataset. The calculated words can be used to enhance the results comparatively better.
关键词:component; Text mining; Sentiment Investigation; ; POS tagging; Sentiwordnet