Consumer Evaluation of Green Foods in China: An Approach from Text Mining and Random Forest on Online Consumer Reviews

Abstract

This study aims to identify consumer evaluation of green-labeled rice in China using e-commerce review data. It also proposes a random forest model to predict consumer evaluation of green-labeled rice. First, using text mining techniques, we summarized the tf-idf scoring for each of the two products reviewed. Second, we constructed a random forest model to find the important words affecting the rating of green-labeled rice. Finally, we used co-occurrence networks to clarify the relationship among keywords that influence consumer evaluations and whether the influence is positive or negative. We found that consumers placed importance on the packaging, texture, taste, price, quality, and likes of green-labeled rice at the time of purchase and after purchase. Moreover, we found that the words green food, traceability, and quality led to good evaluations of green-labeled rice. Chinese consumers were found to be more likely to purchase products with quality certification labels, but it was also found that green food certification is not necessarily an attribute that consumers value most.