Table 1 Detection systems and their features
Researcher | Method | Features | Dataset | F1-score |
|---|---|---|---|---|
McGinnis et al. [35] | Logistic regression and linear SVM | Zero crossing rate, Mel frequency cepstral coefficients and the Z-score of the power spectral density | McGinnis et al. [35] | – |
Tadesse et al. [32] | SVM | LIWC, LDA and Bigram | Pirina and Çöltekin [44] | 0.91 |
Islam et al. [43] | Coarse KNN | LIWC | Islam et al. [43] | 0.71 |
Reece et al. [31] | Random Forest | LIWC, LabMT, ANEW and Unigram | Reece et al. [31] | 0.61 |
Hassan et al. [30] | SVM | N-gram, POS tagger, Sentiment Analyser and Negation | Hassan et al. [30] | 0.81 |
Shen et al. [42] | Multimodal dictionary learning | LIWC, VAD, LDA, word2vec and Twitter behaviour data | Shen et al. [42] | ~ 0.85 |
Deshpande and Rao [29] | Multinominal Naive Bayes | Bag-of-words | Deshpande and Rao [29] | 0.83 |
Tsugawa et al. [33] | SVM | Bag-of-words, LDA, sentiment analysis+user specific information | Tsugawa et al. [33] | 0.46 |
De Choudhury et a.l [39] | SVM | ANEW,LIWC and Twitter behaviour data | De Choudhury et al. [39] | 0.68 |