Deep CNN with SVM-Hybrid Model for Sentence-Based Document Level Sentiment Analysis Using Subjectivity Detection

With the growth of e-commerce reporting, online customer reviews have evolved rapidly, voicing the sentiment or opinion of customers about goods. The analysis of belief could provide useful data for us. Sentiment analysis on social media like Twitter or Facebook, is now the comprehensive way of understanding about the views of customers and has extensive variety of applications. In the context of NLP, automated text classification can be a fundamental activity and it can help people to pick essential information from vast text resources. Sentiment analysis may be a computational technique that plays a key role in automating the retrieval of subjective knowledge, i.e. customer’s sentiment from online text reviews or opinion from social network like Twitter and Facebook. Lexicon-based and machine learning-based methods are two main approaches widely used in sentiment analysis activities. In machine learning based framework, Sentiment analysis is a text recognition task. The outcome depends not only from the soundness of the algorithm for machine learning, but also with the appropriate features. In recent years, the most recent technological advancements, like deep-learning techniques, have resolved a number of standard challenges caused by the lack of lexical tools in the region. It has been exhibited that deep-learning models are auspicious and potential tool to NLP challenges. In this work, the fusion of deep CNN with SVM will automatically detect and extract subjective sentence-level features to perform sentiment analysis of online product review dataset with highest accuracy and less computation time.