Research Product Reviews Emotions Mining Based Attribute

Because of the rapid development of e-commerce,more and more users prefer to buygoods online,and they prefer to buy goods refer to the other users’ reviews to determine ifthe goods suitability or not.Currently,many e-commerce sites have an overall emotionaltendencies analysis of the goods.,but users more likely to choice the goods according to thegoods properties.Study for goods’ reviews,we research the technologies which need to construct thereviews emotions mining system,including text preprocessing、goods properties extractionand association、property level emotional tendency analysis and analysis shows.Connection dictionary algorithm and statistical algorithm to segmentation,and using theViterbi algorithm to find the optimal path of the hidden Markov model to tag. Theexperimental results show that the improved segmentation algorithm can achieve a higherperformance compared to before and achieve a higher precision.Using the improved association rules algorithm and three pruning rules to extractfrequent attributes and association properties words and viewpoint word association,toobtain more faster.Design reviews based on properties of sentiment analysis and rating methods, theemotion in accordance with the degree of positive or negative, we divided the strength intofive levels.The experimental results verify the feasibility of the proposed algorithm andcalculate the accuracy rate...