Table 2.

Results of different techniques.

Models	10-Fold Cross Validation		Independent
Predictor	ACC	F-Score	ACC	F-Score
LSTM-emb	0.700	0.735	0.734	0.779
CNN-emb	0.704	0.739	0.733	0.776
BiLSTM-onehot	0.725	0.729	0.757	0.777
CNN-onehot	0.719	0.731	0.748	0.778
CNN-onehot-PCA	0.748	0.750	0.768	0.786
Comb-emb-PCA (UbiComb)	0.804	0.795	0.818	0.825
RF-Comb	0.762	0.757	0.781	0.800

The Comb-emb-PCA (UbiComb) provided the best results in terms of 10-fold cross-validation and independent results.