Research Of How Online Reviews Impact On User Behavior In Structural Point Of View

With the development of technology and equipment on the Internet,shopping online progress quickly. As the limitation of present product andservice, online word-of-mouth came out in order to help users to know productsbetter. The volume of word-of-mouth increases rapidly. And they shows the realfeeling of users who have used it in deferent ways. The information can be one ofthe most important elements in judging a good product. Which characteristics oftext word-of-mouth will be reflected by quantitative relation of structural onlineword-of-mouth? What will produced by text word-of-mouth after classification,choice and combination? How to further process word-of-mouth using emotionclassification algorithm in real application?After read and organized the literatures of domestic and foreign, wesummarized and analysed them. Combining theory of online word-of-mouth, userbehavior and emotion classification, we research the quantitative relationbetween online word-of-mouth and user behavior. And then we study thisproblem in a practical way. The following is content and innovation of ourresearch.(1) What features of word-of-mouth influence user behavior? Onlineword-of-mouth has different length, number and emotional tendency. We usestatistical analysis to ensure whether these properties have influence on userattention and their purchase. Which property has the largest influence?(2) Thedifference and relation of different style word-of-mouth. We can obtain thestatistical result of structural online word-of-mouth. We try to find thecharacteristic of text word-of-mouth from analysis result. Then we use it todevelop new value of online word-of-mouth.(3) The emotion tendency of onlineword-of-mouth can influence user behavior. We analyze several emotionclassification algorithms. We find the suitable algorithm, N-gram algorithm, forour study. Based on supervised learning of emotion classification, we achieve anaccuracy and efficient classification result. This study realized clause emotionclassification and combined them in an order. It is helpful for merchants to sell.(4) There is important efficient in practical application. Combined with thecharacteristics of online word-of-mouth research and conclusion of user behavior,we deal with it through the classification of emotional way. Using improvedword-of-mouth in telephone application, we learn the different user behaviorafter this.