Recently, reviews of goods are one of the important factors which have an influence on the consumer's buying behavior. Under the situation, information aggregation of reviews has attracted much attention in electronic commerce (EC) site. In this study we propose a review recommendation system target on golf EC site. In our proposal system, reviews are scored by the evaluation of characteristic words which were obtained by TF-IDF and reputation analysis. Moreover, our proposal system visualize concurrent network using concurrent relations between the characteristic words. In order to verify the effectiveness of our proposal system, we conducted an experimental evaluation. In the experimental evaluation, we compared our proposal reviews with other review (selected randomly, recently date and reference counts). As the result, it was found that the result score of our proposed reviews were almost the same as the result score of reference count. Moreover, we received positive evaluation about concurrent network. Through these experimental evaluate, we consider that the effectiveness of this system was successfully verified.
[1]
Takashi Inui,et al.
Structuring Opinions by Relative Characteristics for User-Opinion Aggregation
,
2013
.
[2]
Ryosuke Yamanishi,et al.
Detection of Evaluation Points Using Appearance Frequency and Syntax Features for Overview of Web Review Structure
,
2015
.
[3]
Yuji Matsumoto,et al.
Applying Conditional Random Fields to Japanese Morphological Analysis
,
2004,
EMNLP.
[4]
Weigang Huo,et al.
Topical text network construction based on seed word augmentation
,
2015,
2015 12th International Conference on Fuzzy Systems and Knowledge Discovery (FSKD).