Positive and Negative Arguments in Review Systems: An Approach with Arguments

In this work we study arguments in Amazon.com reviews. We manually extract positive (in favour of purchase) and negative (against it) arguments from each review concerning a selected product. Moreover, we link arguments to the rating score and length of reviews. For instance, we show that negative arguments are quite sparse during the first steps of such social review-process, while positive arguments are more equally distributed. In addition, we connect arguments through attacks and we compute Dung’s extensions to check whether they capture such evolution through time.