ML method | Reference | Objectives | Materials | Output |
---|---|---|---|---|
SVM | [25] | Design opinion classifier for classifying opinions from Bangla text data | Twitter text, English, Bangla | Positive and negative |
SVM | [59] | To extract multi-class emotions from Malayalam text using the proposed approach | Malayalam text | Emotions (joy, sadness, anger, fear, surprise or normal) |
SVM | [60] | To determine the expressed sentiment towards a specified aspect category in a given sentence | Yelp restaurant reviews corpus | Negative, positive and neutral |
SVM | [61] | To propose and analyse new emotion identification method based on online medical knowledge-sharing community | Medical service comments | Positive and negative |
SVM | [26] | To address the challenge of analysing the features of negative sentiment tweets | Twitter (TREC Microblog Track 2013) | Negative |
SVM | [62] | To rank colleges based on a single feature, multiple features and no feature | Twitter (colleges) | Positive, negative or neutral sentiment |
SVM | [27] | To determine the polarity of Facebook comments “positive or negative” | Facebook dataset (Tunisian political pages) | Positive and negative |
SVM + RF | [31] | Determines polarity of reviews given by users and provide recommendation list | Twitter stream | Positive and negative |
SVM + ANN + RF | [7] | To evaluate the thoughts of users in the IMDB movie reviews on tweets obtained from different outlets | IMDB dataset, Review Movie | Positive and negative |
SVM + CRF + Multinomial NB | [53] | To present an ensemble framework of text classification which reviews products | Twitter and product review | Positive and negative |
SVM + NB + RF | [50] | Mining consumer reviews with a machine learning approach by converting reviews into vector representations for classification | Amazon review dataset | Positive or negative |
SVM + Multinomial NB + DNN | [56] | To compare multiple state-of-the-art models capable of classifying game reviews as positive, negative or neutral | Games reviews | Positive, neutral and negative |
NB + ME + SGD + SVM | [47] | To classify human sentiment-based movie reviews using various supervised machine learning algorithms To examine the accuracy of different methods | Internet Movie Database (IMDB) | Positive, negative and neutral |
KNN + SVM + RF | [63] | To classify sentiments into positive, negative or neutral polarity using a new similarity measure | Stanford Twitter dataset | Positive, negative or neutral polarity |
SVM + Multinomial NB + LR + RF | [52] | To develop a clinical decision support system for the personalised therapy process | Drug review dataset | Positive, negative or neutral |
NB + SVM + DBN | [44] | To classify a Malay sentiment by proposing a classification model to improve classification performances | Online blogs and forums of Malaysian website | Positive and negative |
Fuzzy rule + SVM + ME | [64] | Social Media data for decision making to purchase and recommend products online | Twitter text reviews | Positive and negative |