Mining The Feature And Emotional Words From Product Reviews Based On The Part Of Speech And Syntactic Relations

After purchasing the products, the customers usually leave their reviews of the productson the platform provided by the traders. The potential consumers can get information of theproduct to make decisions; the traders can also learn the advantages of other traders fromthose reviews and then correct their own deficiencies. The emotional words in the reviewsplay a decisive role for the analysis of the attitudes, for the feature words described by theemotional words can reflect the focuses of the customers’ attention. One of the most importanttasks during the exploring the product reviews is to set the emotional words and feature wordsas the researching object.This paper works on the following aspects:This paper proposes the preprocessing methods of standardizing and splitting the productreviews to unify the format of the reviews, then divides the reviews into short sentences. Bypreprocessing of the reviews, the effect of processing the reviews is enhanced.This paper proposes the method of exploring the emotional words based on the parts ofspeech, taking the collocation of the emotional words and feature words into consideration.By extracting speech template of matching the emotional words from the seed reviews, thismethod applies this template to explore the emotional words, then prune the results by bothstop word set and dependencies. The final result contains the emotional words that reviewboth the apparent features of the product and implicit features.This paper proposes the method of exploring the feature words on the basis of theemotional words. Which sentence fragment the feature words lie can be positioned accordingto the position of the emotional words in the reviews. Through the combination ofpart-of-speech collocation and syntactic structure, to automatically match the feature wordsand then reach the corresponding of feature words and emotional words. The matching resultscan make good identification of both high-frequency and low-frequency feature words.The experimental result shows that a better exploring effect can be obtained by applyingthe exploring method proposed by this paper.