One-step and Two-step Classification for Abusive Language Detection on Twitter

By Ji Ho Park, Pascale Fung. 2017

380 citations recorded by the source corpus.

Automatic abusive language detection is a difficult but important task for online social media. Our research explores a two-step approach of performing classification on abusive language and then classifying into specific types and compares it with one-step approach of doing one multi-class classification for detecting sexist and racist languages. With a public English Twitter corpus of 20 thousand tweets in the type of sexism and racism, our approach shows a promising performance of 0.827 F-measure by using HybridCNN in one-step and 0.824 F-measure by using logistic regression in two-steps.

Subfields: Artificial Intelligence, Information Systems, Natural Language Processing

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Information Systems

In Natural Language Processing

Citations in the corpus

RealToxicityPrompts: Evaluating Neural Toxic Degeneration in Language Models (2020, background): “We choose these two lexical categories since profanity or swearwords often carry toxicity, and identity mentions are known to co-occur with hateful and toxic language (specifically, minority identity mentions; Park and Fung, 2017; Dixon et al., 2018).”
The Risk of Racial Bias in Hate Speech Detection (2019, background): “Past work on bias in hate speech datasets has exclusively focused on finding and removing bias against explicit identity mentions (e.g., woman, atheist, queer; Park and Fung, 2017; Dixon et al., 2018).”
Reducing Gender Bias in Abusive Language Detection (2018, background): “(Park and Fung, 2017), Gated Recurrent Unit (GRU) (Cho et al.”
Large Scale Crowdsourcing and Characterization of Twitter Abusive Behavior (2018, neutral): “(Chen et al. 2012), (Razavi et al. 2010).”
Challenges for Toxic Comment Classification: An In-Depth Error Analysis (2018, background): “…2017; Warner and Hirschberg, 2012), online harassment (Yin and Davison, 2009; Golbeck et al., 2017), abusive language (Mehdad and Tetreault, 2016; Park and Fung, 2017), cyberbullying (Dad-var et al., 2013; Dinakar et al., 2012; Hee et al., 2015; Zhong et al., 2016) and offensive language (Chen…”
The Linguistic Ideologies of Deep Abusive Language Classification (2018, methodology): “, 1990), can be applied to sequences of word embeddings in a similar manner to how they are applied to bitmap images, and have been shown to perform well in some abusive language detection tasks (Park and Fung, 2017).”
HABERTOR: An Efficient and Effective Deep Hatespeech Detector (2020, background): “, 2019), and proposed deep neural network architectures (Zampieri et al., 2019; Gambäck and Sikdar, 2017; Park and Fung, 2017; Badjatiya et al., 2017; Agrawal and Awekar, 2018).”
DeepHate: Hate Speech Detection via Multi-Faceted Text Representations (2020, methodology): “In learning post’s topic representation, we set the number of topics in WZ-LS, DT, and FOUNTA datasets to 15, 10, 15, respectively.”
Detect All Abuse! Toward Universal Abusive Language Detection Models (2020, methodology): “The reason is that the CNN used in OTH can capture the information for phrases, which is the “white nationalist” here.”
Cross-Domain Detection of Abusive Language Online (2018, methodology): “Some approaches (Badjatiya et al., 2017; Park and Fung, 2017; Mehdad and Tetreault, 2016) also rely on combining different types of models.”
Tackling Online Abuse: A Survey of Automated Abuse Detection Methods (2019, methodology): “Park and Fung (2017) utilized character and word-level CNN s to classify comments in the dataset that they formed by combining DATA - TWITTER -”
Effective hate-speech detection in Twitter data using recurrent neural networks (2018, methodology): “The inferiority of FastText over LSTM is also reported in the work in [1], as well as being inferior over CNN [19].”
Bias in word embeddings (2020, background): “[78] have attempted to create classifiers that detect abusive language.”
Toxicity Detection can be Sensitive to the Conversational Context (2021, methodology): “Then, deep learning techniques were employed, operating on word embeddings (Park and Fung, 2017; Pavlopoulos et al., 2017b,c; Chakrabarty et al., 2019; Badjatiya et al., 2017; Haddad et al., 2020; Ozler et al., 2020).”
Comparative Studies of Detecting Abusive Language on Twitter (2018, methodology): “873 .848 .860 .788 .804 .783 SVM (word) .”
Context Sensitivity Estimation in Toxicity Detection (2021, background): “…has been addressed by machine learning approaches (Davidson et al., 2017; Waseem and Hovy, 2016; Djuric et al., 2015), including deep learning approaches (Park and Fung, 2017; Pavlopoulos et al., 2017b,c; Chakrabarty et al., 2019; Badjatiya et al., 2017; Haddad et al., 2020; Ozler et al., 2020).”