The features-based evolution of the opinions in forums using a semantic summarization approach

Abstract

The consumers do not often express their opinions on the entity itself, but about its features. Our overall goal is represented by the visualization of the evolution of opinions in a collection of forums data, which simplifies the graph with messages that contain opinions. In the final graph, each node resulted from the aggregation of the opinions with the same polarity from the neighboring nodes contains the opinion polarity and words that give the respective polarity. For this purpose, in this paper we identify the features for a target entity, we search opinions about features and we analyze the opinion sentences. In the proposed solution, we adopt an approach to opinion summarization.