In order to help merchants understand consumers'demand preferences for commodity and the composition of con-sumer bases,product demand preferences in online comments and customer group models are constructed based on dictionary-di-vided sentiment analysis and K-means clustering.The online comments of Huawei Mate60 series mobile phones from the Jingdong platform are crawled and then processed.The LDA topic model is used to determine consumers'topic of interest,and sentiment analysis using HowNet dictionary combined with custom dictionary is used to calculate sentiment scores.Finally,consumer segmen-tation is obtained based on word vectors and K-means clustering algorithm.This can help merchants develop clear product position-ing and characteristics according to the characteristics of different clustering groups to meet consumer demands.
万方数据