According to the characteristics of existing online reputation system,we divide the sentiment intensity into several levels.Considering the fuzziness of sentiment intensity,we set up membership functions for each sentiment intensity level.We then develop a method based on fuzzy-statistics of sentiment words to obtain sentiment intensity of online reviews.Fuzzy-statistic experiment is designed to acquire the membership of sentiment words in online reviews on mobile phones.The experimental results show that the membership value of the sentiment words are stable and centralized; negative words not only change the polarity of sentiment words but also weaken sentiment intensity of the words; degree adverbs strength the sentiment intensity of sentiment words,but the degree of strengthening decreases with the increasing intensity.
国家科技期刊平台
NSTL
维普
万方数据
