Abstract

Trustworthiness of an e-vendor in the e-marketplace can be determined in multiple dimensions: product, price, and shipping. An e-vendor who has high trust level in more dimensions is more likely to have the competitive advantage than others. A consumer’s review is analyzed to find its polarity in different dimensions. Positive sentiment in consumers’ reviews helps increase the trustworthiness of e-vendors which in turn influences consumer’s purchase intention. In this paper, we propose the method to automatically analyze Thai sentiment of consumer’s review in product, price, and shipping dimensions
by using multi-dimensional lexicon and sentiment compensation technique. A consumer’s review in Thai language is tokenized using the longest matching algorithm. Then, it is analyzed to find its sentiment. Sentiment compensation technique is used to automatically compensate the sentiment to a dimension where consumer’s review mentions the sentiment without a dimension. The results show that our proposed method outperform sentiment to dimension (S2D) and dimension to sentiment (D2S) methods with the overall accuracy 93.60%.

Index Terms—multi-dimensional trust, multi-dimensional lexicon,
Thai sentiment analysis, sentiment compensation, e-marketplace