Classification of Sentiments on online products using Deep Learning Model – RNN
Lakshmidevi N1, M. Vamsikrishna2, S. S. Nayak3
1Lakshmidevi N, Research Scholar and Assistant Professor Centurion University of Technology and Management , Paralakhemundi, and GMR Institute of Technology, Rajam, India.
2Dr. M. Vamsikrishna, Professor, CSE, Chaitanya Institute of Science and Technology, Kakinada, India.
3Dr.S.S.Nayak, Professor, Centurion University of Technology and Management, Paralakhemundi, India.
Manuscript received on September 23, 2019. | Revised Manuscript received on October 15, 2019. | Manuscript published on October 30, 2019. | PP:7165-7172 | Volume-9 Issue-1, October 2019 | Retrieval Number: A1910109119/2019©BEIESP | DOI: 10.35940/ijeat.A1910.109119
Open Access | Ethics and Policies | Cite | Mendeley
© The Authors. Blue Eyes Intelligence Engineering and Sciences Publication (BEIESP). This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/)
Abstract: Due to advancement of technology there is a large usage of social media which leads to demand for data in the web. This data is very helpful to categorize the opinions into different sentiments and general evaluating the mood of public. The current research contributions are towards to detect the complete separation of sentence regardless of their aspects. The computational observation of sentiments and opinions stated by people in written language. Examination of defies presented by informal and crisp micro blogging created the origins. The proposed work targets building up a model for conclusion characterization that investigates the product features. It is also addresses domain explicit vocabularies to offer an domain arranged methodology and subsequently dissect and extricate the purchaser opinion towards well known advanced cell marks in the course of recent years. This model describes the use of deep learning model such as recurrent neural network to get better accuracy over traditional machine learning methods such as Random forest, Naïve bayes. The RNN model got training accuracy 97.6% and testing accuracy 95.6% which are much better compared to traditional machine learning models.
Keywords: Sentiment Analysis, SVM, Opinion mining, Recurrent Neural Network, Natural Language Processing, Naive Bayes, Random Forest, Product Reviews.