|
| Authors | Features used | Classifier |
|
| Nandhini and Sheeba [20] | Noun, pronoun, and adjective | Fuzzy logic-based genetic algorithm |
| Potha et al. [21] | Local, sentimental, contextual, and gender-specific language features | SVM |
| Kumar and Sachdeva [28] | Direct and indirect CB features | SVM |
| Al-garadi et al. [8] | Network, activity and user information, and tweet content | SVM |
| [28] | Network, activity and user information, and tweet content | Naïve Bayes (NB) |
| [25] | Network, activity and user information, and tweet content | k-nearest neighbor (KNN) and random forest (RF) |
| Balakrishnan et al. [25] | Psychological features | NB, RF, and J48 |
| Murnion et al. [18] | IsAbusive, IsPositive, IsNegative, HasBadLanguage, IsRacist, NoobRelated, SpecificTarget, and FilteredText | Sentiment text analytics system is supported with a scoring scheme |
| Ho et al. [27] | Abusive words | Logistic regression model |
| Balakrishnan et al. [24] | 15 twitter features [23] | RF classifier |
| Sánchez-Medina et al. [26] | Psychopathy, narcissism, and machiavellianism | Ensemble classification trees |
| Lee et al. [22] | New abusive words | Three-layered neural network model |
|
