Ranking Online Customer Reviews with the SVR Model

On the online E-Commerce platform, customer reviews provides valuable opinions and relevant content, which will affect the perches behavior of other customers. Since the amount of online review grow fast, it is hard to read them all, therefore, a system that can find the reviews with better quality is necessary. In order to better understand the quality of reviews. In this paper, we proposed a system that can rank the reviews based on a set of linguistic features and a Support vector regression (SVR) model as a scorer. To evaluate our system, we collect 3730 Chinese reviews in eight product categories (books, digital cameras, tablet PC, backpacks, movies, men shoes, toys and cell phones) from Amazon.cn with the voting result of whether the review is helpful or not. Since the voting result might be biased by voting time and total voting number. We defined 4 types of evaluation index and compare the regression result to each index.