Table 3.
Predictive models and their primary metrics observed in reviewed studies.
| Predictive model used | ||||||||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Data type | No. | Ada-Boost | CNN | GRU | KNN | LR | LSTM | MLP | NB | Proposed Algo | RF | DT | SVM | XGBOOST | Others | Primary Performance metrices | Performance range(for data type) | Language of Data Samples Analyzed | Range of participants | References |
| 18 | Y | Y | Y | Y | Y | Y | Y | Y | Y | Y | Y | BRR, NLP,SVR | F1 score: 7 Accuracy: 8 RMSE: 1 Not used:1 | 0.47 to 0.99 | English - 14 Bangla-1 Indonesian- 1 Spanish- 1 | 55 - 20,000 | S2,S9,S12,S13, S17,S21,S22,S31,S34,S35,S36,S42,S45,S46,S47, S48,S50,S52 | |||
| 6 | Y | Y | Y | Y | Y | GPR | F1 score: 3 Accuracy: 3 | 0.61 to 0.90 | English - 6 | 90 - 5,947 | S11,S16,S20,S27,S42,S43 | |||||||||
| 2 | Y | lr | Accuracy: 1 AUC:1 | 0.60 to 0.71 | English - 2 | 749 - 1,908 | S14,S28 | |||||||||||||
| Reditt | 13 | Y | Y | Y | Y | Y | Y | Y | Y | Y | Y | Y | F1 score:10 Accuracy: 2 ERDE: 1 | 0.043 to 0.96 | English -13 | 365 - 48,537 | S1,S3,S7,S8,S19,S23, S24,S25,S33,S41,S44,S49,S54 | |||
| Sina Weibo | 6 | Y | Y | Y | Bert,MFFN, built models based on linguistic and behavioral features | F1 score: 3 Accuracy: 2 Precision-:1 | 0.5-0.97 | English - 3 Chinese- 3 | 1000 - 30,000 | S4,S6,S10,S30,S37, S39 | ||||||||||
| facebook + twitter | 4 | Y | Y | Y | Y | Y | VGG-Net | F1 score: 1 Accuracy: 2 Pearson correlation: 1 | 0.56 to 0.77 | English - 4 Bangla- 1 | 150 - 3,498 | S15,S32,S38,S53 | ||||||||
| others | 5 | Y | Y | Y | Y | Y | GBM,Multi-class tree models | F1 score: 3 Accuracy: 1 Recall: 1 | 0.86 to 0.96 | English - 5 | 619 - 1,000,500 | S5,S18,S26,S31,S51 | ||||||||