|
Research work | Publication year | Source of data | Content analysis | Contextual analysis | Learning algorithm class | Reference word list sources | Dictionary/word list generation proposal |
|
Agrewal and Awekar [26] | 2018 | FormSpring, Twitter, Wikipedia | Yes | No | Deep neural network, support vector machine, logistic regression, random forest, Naïve Bayes | None | No |
Aind et al. [27] | 2020 | Multiple publicly available datasets | Yes | No | Novel algorithm based on reinforcement learning | GitHub reference wordlists for profanity and sentiment dictionaries (refer to paper) | No |
Balakrishnan et al. [28] | 2020 | Twitter | Yes | Yes | Random forest, decision tree | Unspecified | No |
Banerjee et al. [29] | 2019 | Twitter | Yes | No | Deep neural network | None | No |
Cheng et al. [30] | 2019 | FormSpring, Twitter | Yes | No | Random forest, Extratree, AdaBoost | Unspecified | No |
Cheng et al. [31] | 2019 | Instagram | Yes | Yes | Novel algorithm (hierarchical attention networks for cyberbullying detection) | Unspecified | No |
Dadvar et al. [32] | 2013 | YouTube | Yes | Yes | Support vector machine | Noswearing website | No |
Dani et al. [33] | 2017 | Twitter, MySpace | Yes | Yes | Linear regression, sparse learning, support vector machine | Unspecified | No |
Dinakar et al. [34] | 2011 | YouTube | Yes | No | Naïve Bayes, rule-based JRip, decision tree, support vector machine | Unspecified | No |
Hosseinmardi et al. [35] | 2015 | Instagram | Yes | Yes | Statistical analysis | Unspecified | No |
Hosseinmardi et al. [36] | 2014 | Instagram, Ask.fm | Yes | Yes | Statistical analysis | Unspecified | No |
Iwendi et al. [37] | 2020 | Kaggle dataset from Facebook, Twitter, Instagram | Yes | No | Deep learning models | None | No |
Kontostathis [17] | 2009 | Perverted-Justice | Yes | No | Decision tree, K-mean clustering | Predation dictionary (refer to paper) | No |
Kontostathis et al. [18] | 2012 | Perverted-Justice | Yes | No | Decision tree, rule-based classifier | Predation dictionary (refer to paper) | No |
Kontostathis et al. [19] | 2013 | FormSpring | Yes | Yes | Essential dimensions for LSI | Noswearing website | No |
Lu et al. [38] | 2020 | Chinese Weibo, Twitter | Yes | No | Convolutional neural network | None | No |
McGhee et al. [12] | 2011 | Perverted-Justice | Yes | No | Decision tree, rule-based classifier, K-nearest neighbour | Predation dictionary (refer to paper) | No |
Nahar et al. [39] | 2014 | MySpace, Kongregate, Slashdot | Yes | Yes | Fuzzy C-mean clustering, fuzzy support vector machine | Unspecified | No |
Ptaszynski [40] | 2019 | Multiple unofficial school websites and forums (see paper for more information) | Yes | No | Novel brute-force pattern extraction algorithm | None | No |
Rafiq et al. [41] | 2018 | Vine | Yes | No | AdaBoost, logistic regression, incremental classifier | None | No |
Raisi and Huang [42] | 2018 | Twitter, Ask.fm, Instagram | Yes | Yes | Novel participant vocabulary consistency | Noswearing website | No |
Renolds et al. [14] | 2011 | FormSpring | Yes | No | Decision tree, rule-based classifier, support vector machine, K-nearest neighbour | Noswearing website | No |
Tahmasbi and Rastegari [43] | 2018 | Twitter | Yes | Yes | Decision tree, rule-based classifier, support vector machine, logistic regression, AdaBoost, Naïve Bayes | Unspecified | No |
Van Hee et al. [44] | 2018 | Ask.fm | Yes | No | Support vector machine | Google profanity list | No |
Wang et al. [45] | 2020 | Instagram, Vine | Yes | No | Novel multimodal cyberbullying detection framework (based on neural network) | None | No |
Xu et al. [46] | 2012 | Twitter | Yes | Yes | Logistic regression, support vector machine, Naïve Bayes, latent topic models | None | No |
Yao et al. [47] | 2019 | Instagram | Yes | No | Novel sequential hypothesis testing model CONciSE | Noswearing website | No |
Yin et al. [15] | 2009 | MySpace, Kongregate, Slashdot | Yes | Yes | Support vector machine | Noswearing website | No |
Zhao et al. [48] | 2020 | Twitter | Yes | No | Support vector machine, logistic regression, random forest, and multiple deep learning models | None | No |
Zhong et al. [49] | 2016 | Instagram | Yes | Yes | Support vector machine, convolutional neural network, deep learning models | None | No |
Gencoglu [50] | 2021 | Jigsaw, Twitter, WikiDetox, Gab Hate Corpus | Yes | Yes | Deep neural network | None | No |
Cheng et al. [51] | 2020 | Instagram, Vine | Yes | Yes | Unsupervised Gaussian mixture model | Unspecified | No |
Kumar and Sachdeva [52] | 2021 | YouTube, Instagram, Twitter | Yes | No | Convolutional neural network, deep neural network | None | No |
Dadvar and Eckert [53] | 2020 | FormSpring, Wikipedia, Twitter, YouTube, | Yes | No | Deep neural networks | None | No |
Wang et al. [54] | 2020 | FormSpring, Twitter | Yes | No | Word2Vec, word similarity scheme | Noswearing website | No |
Fang et al. [55] | 2021 | Twitter, Wikipedia | Yes | No | Neural network with gated recurrent unit | None | No |
Rezvani et al. [56] | 2020 | Instagram, Twitter | Yes | Yes | Neural network | Google profanity list | No |
Current work | | YouTube + FormSpring | Yes | No | Decision tree, Naïve Bayes, rule-based classifiers | Noswearing website + generated dictionary | Yes |
|