Customer Review Behavior Of Smart Phone Users:A Cross Cultural Study Based On Sentiment Analysis

With the development of internet technology,the internet has been regarded as an important channel for customer complaints.It becomes a challenge in user centered design how to analyze the customer complaint big data from customers with different cultural background,to identify the customer behaviors behind these data and cross-cultural differences between countries.In this study,a methodology based on Natural Language Processing(NLP)was adopted for analyzing cross-cultural differences and similarities,using customer reviews on a particular smart phone brand on the internet as data source and taking India and North America(United States and Canada)as examples.Data from social media,e-commerce sites,message boards of service centers,and some product review sites was preprocessed and vectorized using NLP techniques.Cultural features,including hot topics,service related score,recommendation related score,text sentiment orientation classification,text emotion intensity classification,and emotion intensity score,were extracted from the corpus by means of Word2Vec,Latent Dirichlet Allocation(LDA),Gradient Boosting Decision Tree(GBDT),and some lexicon based methods.It was hypothesized that cultural difference has effects on user complaint behaviors at both topic level and emotion expression level.The results of this study suggest that North American customers concern more about service than Indian customers in average while the service related reviews may be more focused in several topics in India.And compared with the North American customers,India customers trend to give recommendations or suggestions to others in their reviews.Neutral reviews in North America are more than in India significantly.As for emotion expression,India customers trend to be more emotional both in their negative and positive reviews.