Attention Mechanism with Gated Recurrent Unit Using Convolutional Neural Network for Aspect Level Opinion Mining

Deep neural network models are emerging in the area of natural language processing and have become a topic of interest in sentiment analysis. The participation of more social media users provides increased information which has made analysis challenging. Aspect level sentiment analysis is used in the identification of the sentiment polarity of a text in different aspects. This paper presents four deep neural network-based methods with varied input word vector representation for the aspect level opinion mining. A novel approach using an attention mechanism with a gated recurrent unit and a convolutional neural network for aspect level opinion mining with different input vector representations is proposed. This work is an addition to the existing research that includes novel approaches for the assessment of the quality of services based on customer reviews in the restaurant domain. Data accumulated on restaurant opinion have been chosen for experimental study, and the results obtained indicate achievement of good accuracy, precision, recall, and f-measure compared to other approaches.