Table 2 Deep learning and neural networks
Researcher | Deep learning architecture | Feature types | Dataset | F1-score |
|---|---|---|---|---|
Kabir et al. [58] | BERT, DistilBERT | BERT | DEEPTWEET [58] | Â |
Ansari et al. [59] | LSTM with Attention | GLoVE, SenticNet | Reddit, CLPsych 2015, eRisk Dataset | 0.77 |
Wani et al. [60] | CNN, LSTM | Word2Vec, TF-IDF | Wani et al. [60] | 0.99 |
Nemesure et al. [61] | Stacked ensemble | Electronic health records; demographic and medical | Nemesure et al. [61] | – |
Zogan et al. [62] | CNN, BiGRU | BERT | Shen et al. [42] | 0.91 |
Wan et al. [63] | Hybrid EEGNet | Resting state EEG | Wan et al. [63] | 0.95 |
Ray et al. [37] | BiLSTM | Audio, text and visual | DIAC [56] | – |
Rosa et al. [53] | CNN, BiLSTM and RNN with SoftMax | – | Rosa et al. [53] | 0.92 |
Tadesse et al. [32] | MLP | LIWC, LDA and Bigram | Pirina and Çöltekin [44] | 0.91 |
Tasnim and Stroulia [36] | DNN | Audio | AVEC ’17 [64] | 0.61 |
Alhanai et al. [34] | LSTM | Audio and text | DIAC [56] | 0.77 |
Cong et al. [49] | XGBoost and attentional-BiLSTM | – | Yates et al. [55] | 0.60 |
Chen et al. [57] | LSTM | – | Chen et al. [57] | – |
Yang et al. [38] | Deep CNN and DNN | Audio and video | AVEC ’17 [64] | – |