| Abstract [eng] |
Consumers experience about goods and services submited to online consumer reviews, helps others consumers to make decision of buying them. In this research reviews performance of online reviews in terms of helpfulness. Using negative binomial regresion found that reviews, with negative sentiment receive less helpfulness. Moreover, the restauran star raiting, review length, frequency and number of a review, positively influence rewiew helpfulness. The most infulencing predictors of review helpfulness are restauran star raiting and negative sentiment of review. Also, reviews was classified in to helpful and not helpful. Classification experimented with artificial neural networks, random forests, k-nearest neighbors models and found that artificial neural networks produced better results. |
