Mining Hot Topics from Free-text Customer Reviews an CRP-based Approach
Huiting Luo · Technology and Innovation Management · 2012
Because people are used to expressing their feelings about the products on internet,comments are many and complex.This study examines how the Chinese Restaurant Process(CRP) model can be used to identify hot topics from free-text customer reviews.This method overcomes the shortcomings that LDA model can't automatically identify the number of hot topics from free-text customer reviews.CRP model can autonomously discover the hot reviews and the updated hot comments,and can accurately identify the hot topics.26,157 restaurant reviews are collected,and the analytical results show that this approach could produce satisfactory results.