Hybrid approach to extract adjectives for implicit aspect identification in opinion mining

Abstract

Opinion Mining (or Sentiment analysis) is a rapidly growing research field that has attracted both academia and industry because of the challenging research problems it poses and the potential benefits it can provide in many real life applications. Aspect-based sentiment analysis, in particular, is one of the fundamental challenges within this research field. In this paper, we propose an approach that combines a corpus based and WordNet (WN) dictionary based models for Implicit Aspect Terms (IAT) extraction. Then, it uses its hybrid model to support Naive Bayes Training for identifying Implicit Aspects. Our schema is structured in three major phases: In Implicit Aspect Representation stage, our approach aims at representing implicit aspects implied by corpus adjectives combined with their WN related words. In Post training stage, the learnt model (on implicit aspect identification) is improved by removing the noisy WN related words. In Implicit Aspect Identification step, we use NB classifier which identifies the final implicit aspects. We conducted experiments using NB classifier on electronic products and restaurant datasets. The results show that : (1) NB performs better using our hybrid model than using corpus based model alone. (2) For more efficiency, our approach needs to adjust the learnt NB model by removing noisy WN terms which negatively bias NB training.