An ensemble approach to stabilize the features for multi-domain sentiment analysis using supervised machine learning

Journal of Big Data

Table 8 Comparisons of performance of proposed approach with different literature using different domain review Dataset

	Dataset	Feature selection method	Classifier	Performance
Pang et al. [29]	Internet Movie Database (IMDb)	N-gram features	SVM NB ME	82.9 (Accuracy) 81.5 81.0
Agarwal et al. [1]	Movie (IMDb) Product (book, DVD, electronics)	N-gram, IG, RSAR, Hybrid(IG + RSAR)	SVM NB	87.7 (F measure) 80.9
Al-Moslmi et al. [44]	Movie reviews in the Malay language	IG, CHI, Gini Index	SVM NB KNN	85.33 (F-measure) 80.88 74.68
Kolog et al. [18]	Sentiment from social network regarding student’s life	N gram features	SMO MNB J48	80.0 83.0 69.0
Tripathy et al. [22, 43]	Movie (IMDb)	N-gram features	SVM ME NB SGD	88.94 88.48 86.23 85.11
Our approach	Movie (IMDb) Electronics product Kitchenware	N-gram, Combination of Unigram and bigram with IG, CHI, Gini Index	SMO MNB RF LR	90.18 (F-measure) 88.18 87.73 87.32