Sentiment Classification using Neural Network and Ensemble Model based on Genetic Algorithm
Kalaivani P1, Logeshwari D2, Tamizhselvi A3

1Kalaivani P*, Department of IT, St.Joseph’s College of Engineering, OMR, Chennai.
2Logeshwari D, Department IT, St.Joseph’s College of Engineering, OMR, Chennai.
3Tamizhselvi A, Department IT, St.Joseph’s College of Engineering, OMR, Chennai.
Manuscript received on May 06, 2020. | Revised Manuscript received on May 15, 2020. | Manuscript published on June 30, 2020. | PP: 1885-1891 | Volume-9 Issue-5, June 2020. | Retrieval Number: B3677129219/2020©BEIESP | DOI: 10.35940/ijeat.B3677.029320
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 (

Abstract: The fast development of web sites and the number of product on these websites are available. The purpose of classification of sentiment is to efficiently identify opinion expressed in text. This paper compares three different optimized models including genetic optimized feature selection method, Genetic Algorithm (GA), ensemble approach that uses information gain and genetic algorithm as feature selection methods incorporated SVM model, Genetic Bagging (GB) and the next method uses optimized feature selection as feature selection technique incorporated back propagation model, Genetic Neural Network (GNN) models are compared. We are tested in sentiment analysis using sample multi-domain review datasets and movie review dataset.. These approaches are tested using various quality metrics and the results show that the Genetic Bagging (GB) technique outperforms in classifying the sentiment of the multi domain reviews and movie reviews. An empirical analysis is performed to compare the level of importance of the classifiers GB, GNN methods with McNemar’s statistical method.
Keywords: sentiment classification, machine learning, feature selection, review, information gain, genetic algorithm, ensemble method, back propagation model.