Change Mining Of User Product Attribute Preference Based On Association Rules

With the development and maturity of interactive network and e-commerce.the number of user reviews on products is growing rapidly.Online product reviews have gradually attracted the attention of e-commerce sellers,product enterprises and peer competitors.Online reviews provide information about product quality,which plays an important role in helping consumers to judge product quality,reduce the uncertainty of purchase decisions and help them make better purchase decisions.In order to better tap the value of online reviews,previous studies extract product attributes from reviews to obtain key features,to obtain the needs of consumers,and provide support for the development of enterprise's product strategy.However,the research on product attributes lacks dynamics,and does not consider the inconsistency between user sentiment and rating.Therefore,this paper starts with online product reviews,replaces user ratings with emotional scores of reviews,and proposes a method of mining user product attributes preferences based on Association rules,which can provide theoretical support for enterprises to better understand user needs,improve research and development efficiency,maintain and develop enterprise advantages,and enhance enterprise competitiveness.This research chooses ZOL product online comment as the research platform,takes OPPO mobile phone R15 series product online comment as an example,uses the quantitative analysis method,uses the web crawler to capture the user's comment data on the product.By processing the grabbed data,using association rules and emotional score method to analyze the extracted product attributes.By mining the trend of association rules in two different periods,comparing the changes of key attributes affecting user satisfaction,the following conclusions can be drawn:(1)Combining association rules and emotional score,product online reviews can be included.Product attributes are analyzed more accurately to avoid inconsistencies between user ratings and actual satisfaction.(2)By mining the trend of association rules based on product attributes,we can compare the association rules in two different periods,identify four change patterns,find out the change direction of key attributes of products,and obtain the change of users' preferences for product attributes.Theoretically,this paper divides user satisfaction by emotional score,uses the method of association rule mining based on product attributes,and finds out the important attributes that affect user satisfaction,which can effectively improve the accuracy of association rule mining.In addition,the trend analysis of association rules in the two periods of Chinese online product reviews makes up for the deficiency of dynamic analysis of association rules in the field of Chinese text.In practice,the results of this study can enable businesses to better understand user needs and provide reference in optimizing product attribute,saving R&D costs and improving sales volume.