Research On The Prediction Of Consumer Buying Behavior For Online Reviews

With the rapid development of the network,consumers can obtain more resource information,and merchants can use more abundant marketing methods and modes.At the same time,it is more difficult for consumers to distinguish and screen information.How to stand out in diversified marketing and attract more consumers is the main problem at present.The consumer buying behavior is a influence factors of dynamic change,is also a can be the factors affecting the change of the consumer purchasing behavior in the process of formation,existing external stimuli including business marketing,and other consumer post-purchase evaluation,etc.,there are internal stimuli,such as consumer preferences,"big data" era constantly pounding the stimulus,make consumer behavior research is more complicated.This paper takes online comments as the research object,deeply explores the affective computing method and dynamic recommendation method based on online comments,so as to study the influence of factors such as affective computing and dynamic recommendation ranking on consumers' purchasing behavior,and obtain the prediction results of consumers' purchasing behavior.Firstly,a fuzzy comprehensive evaluation method w-fce based on Word2 vec is proposed to solve the emotion calculation problem of online comments.Word2 vec is used to transform the text comment into a vector representation that is easy to calculate,and word position spacing is used as a rule to match the evaluation object and evaluation emotion information contained in the comment,and then the fuzzy comprehensive evaluation method is used for emotion calculation.Secondly,considering the dynamic nature of online comments and user preferences,the w-fce method is used to calculate the characteristic emotion evaluation,the overall emotion evaluation and the user preference of commodities.Finally,based on the results of emotion calculation and dynamic recommendation ranking,the influence of comment content,emotion characteristics and product recommendation ranking on consumers' purchasing behavior was studied.Based on bayesian network to predict consumer buying behavior,and on the Meituan network hotel review data for empirical analysis.The results of empirical analysis show that the influencing factors of consumers' purchasing behavior are listed in the order of influence degree from large to small: dynamic recommendation ranking > comment score > comment quantity of emotion > comment.Among these four factors,consumers with the highest purchase probability recommended the top ten products,while consumers with the highest purchase probability fluctuated the comment sentiment score.Separately,as the number of comments increases,the probability of purchase behavior decreases first and then increases.With the increase of comment sentiment score,the probability of purchase behavior increased first and then decreased.With the increase of comment score,the probability of purchase behavior is increasing.With the increase of recommendation ranking,the probability of purchase behavior is decreasing.Therefore,merchants can give priority to adjust the marketing strategy of recommendation ranking factor and comment on the emotional factor.