Research on Topic Mining of Breastfeeding Bra Comments Based on K-means and Word2vec
Objective In order to understand the focus of consumers when purchasing breastfeeding bras on on-line platforms,the article constructs breastfeeding bra themes by extracting effective keywords from online reviews,and assists merchants in understanding the direction of consumers'focus by calculating the importance of the themes.Methods The TF-IDF keyword extraction algorithm is selected,combined with K-means and Word2vec for se-mantic clustering,topic identification,topic word mining and topic importance calculation.Results The results showed that the ranking of topic importance after clustering breast feeding bra review texts was:product quality(45.47%),product appearance(35.83%),and product service(18.79%).Conclusion The results confirm that the method can effectively identify and construct the breastfeeding bra theme and theme words,and at the same time,through the importance of the theme,it can understand the key direction of consumers'concern for the purchase of breastfeeding bra on the online platform,which provides theoretical references for breastfeeding underwear enterprises to improve their products and production and soon.
NETL
NSTL
万方数据
