Topic Modeling to Extract Information from Nutraceutical Product Reviews

Deena John, Ernest Kim, Kunal Kotian, Ker Yu Ong, Tyler White, Luba Gloukhova, Nicholas Ross and Diane Woodbridge

Abstract:

Consumer purchases of Vitamins and other Nutraceuticals have grown over the past few years with most of the growth occurring in on-line purchases. However, general e- commerce platforms, such as Amazon, fail to cater to consumers’ specific needs when making such purchases. In this study, the authors design and develop a system to provide tailored information to consumers within this retail vertical. Specifically, the system uses Natural Language Processing (NLP) techniques to extract information from user-submitted nutraceutical product reviews. Using Natural Language Processing, three information streams are presented to consumers (1) a five point rating system for cost, efficacy and service, (2) a summary of topics commonly discussed about the product and, (3) representative reviews of the product. By presenting product-specific information in this manner we believe that consumers will make better product choices.

Paper presented at the 2019 IEEE CCNC Conference.

link here