Extracting Opinion Relations from Online Reviews Based on WAM

Data mining process of pattern discovering in large data sets in that the emerging field is sentiment analysis. Opinion mining is the study of analyzing the human‘s opinions, sentiments, and emotion towards the entities such as products, services. The main application of sentiment analysis is collecting the online reviews about the product, social networks informal text. The process of the opinion targets and the opinion words extraction and determining the relations between these words. In previous, the nearest neighbor rules approach was used, the disadvantage of this method was not suitable for long span sentences. The Word Alignment Model is proposed to extract the opinion words and opinion targets from the obtained reviews and the graph based co-ranking algorithm is used to detect the opinion relation through opinion relation graph. While compared with previous used methods, this novel approach effectively decreases the error probability. The results shows the algorithm effectively outperforms when compare to existing methods.