Cyberbullying Detection with a Pronunciation Based Convolutional Neural Network

Zhang, Xiang; Tong, Jonathan; Vishwamitra, Nishant; Whittaker, Elizabeth; Mazer, Joseph P.; Kowalski, Robin M.; Hu, Hongxin; Luo, Feng; Macbeth, Jamie; Dillon, Edward

doi:10.1109/icmla.2016.0132

Cited by 91 publications

(58 citation statements)

References 17 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Some works observed that the content information in social media has many incorrect spellings, and in some cases, the users in social media intentionally obfuscate the words or phrases in the sentence to evade the manual and automatic detection. 23,24 These extra words will expand the vocabulary and affect the various performances of the algorithm. Waseem and Hovy performed a grid search over all possible feature set combinations.…”

Section: Cyberbullying Detectionmentioning

confidence: 99%

Cyberbullying detection in social media text based on character‐level convolutional neural network with shortcuts

Zhang

et al. 2020

Concurrency and Computation

View full text Add to dashboard Cite

As people spend increasingly more time on social networks, cyberbullying has become a social problem that needs to be solved by machine learning methods. Our research focuses on textual cyberbullying detection because text is the most common form of social media. However, the content information in social media is short, noisy, and unstructured with incorrect spellings and symbols, and this impacts the performance of some traditional machine learning methods based on vocabulary knowledge. For this reason, we propose a Char-CNNS (Character-level Convolutional Neural Network with Shortcuts) model to identify whether the text in social media contains cyberbullying. We use characters as the smallest unit of learning, enabling the model to overcome spelling errors and intentional obfuscation in real-world corpora. Shortcuts are utilized to stitch different levels of features to learn more granular bullying signals, and a focal loss function is adopted to overcome the class imbalance problem. We also provide a new Chinese Weibo comment dataset specifically for cyberbullying detection, and experiments are performed on both the Chinese Weibo dataset and the English Tweet dataset. The experimental results show that our approach is competitive with state-of-the-art techniques on cyberbullying detection task.

show abstract

Section: Cyberbullying Detectionmentioning

confidence: 99%

Cyberbullying detection in social media text based on character‐level convolutional neural network with shortcuts

Zhang

et al. 2020

Concurrency and Computation

View full text Add to dashboard Cite

show abstract

“…It helped to develop gender based approach to classify discrimination which was a type used for bullying [10] [11]. Another type of detecting cyberbullying is by using deep learning and by using neural networks to reduce the problem of unwanted data and bullying data [6]. Then later with increase in use of local abusive language they further developed supervised classification techniques (algorithms that learn from labelled data) with categorizing text data by NLP (Natural Language Processing) feature to analyze the data which was better than deep learning approach [7].…”

Section: Related Workmentioning

confidence: 99%

Behavioral Analysis of User Data on Social Media Applications using Machine Learning Algorithms

C¹

2020

IJMTST

View full text Add to dashboard Cite

This paper is developed to predict user behaviour as cyber bullying on social media by using machine learning algorithms in the big data era. In online platform the use of electronic applications to bully a person by sending messages in the form of aggressive behavior of intimidating or threating nature is termed as cyber bullying. These social networks are the result of long process that began around the 1960's which created the first public network. Social media brings together everyone in an online community where one can communicate, share, post, explore, or stay in touch with one another. We can simply say that social media has changed the way we live. Along with the benefits on one side we can't ignore the facts that social media also pose a serious threat to human. In this paper these threats are considered as human aggressive behavior and we have classified it as bullying words and are categorized as filters which include sections like threats, offensive, abusive, harassment, sexual, violence and so on. The use of these social media has been one of the reason for the generation in huge amount of data and in order to manage and solve real-world problems we use big data along with ML. The machine learning techniques like text-classification approach and lexicon based model are used to categorize the collected data from user and detect the presence of cyberbullying activity. Once the bullying activity is detected we can block that comment automatically which then sends alert message to bully and also to cyber-victim, and also keep the count on number of times the bully words are used. As a result, the admin who has all the privileges can monitor the bullying activity based on the count and can take further action whereas the application will have already blocked the comment which contained the bully word thus leaving no harm to victim. The work presented here has profound implications for future studies of cyberbullying and may one-day help to solve the problem of cyberbullying in social media and other online platforms.

show abstract

“…Another type of approaches using Deep Learning and Neural Networks. One of the proposed methods is Zhang et al [11] in their paper uses novel pronunciation based convolution neural network (PCNN), thereby alleviating the problem of noise and bullying data sparsity to overcome class imbalance. 1313 messages from twitter, 13,000 messages from formspring.me.…”

Section: Related Workmentioning

confidence: 99%

Social Media Cyberbullying Detection using Machine Learning

Hani¹,

Mohamed²,

Mostafa³

et al. 2019

IJACSA

View full text Add to dashboard Cite

With the exponential increase of social media users, cyberbullying has been emerged as a form of bullying through electronic messages. Social networks provides a rich environment for bullies to uses these networks as vulnerable to attacks against victims. Given the consequences of cyberbullying on victims, it is necessary to find suitable actions to detect and prevent it. Machine learning can be helpful to detect language patterns of the bullies and hence can generate a model to automatically detect cyberbullying actions. This paper proposes a supervised machine learning approach for detecting and preventing cyberbullying. Several classifiers are used to train and recognize bullying actions. The evaluation of the proposed approach on cyberbullying dataset shows that Neural Network performs better and achieves accuracy of 92.8% and SVM achieves 90.3. Also, NN outperforms other classifiers of similar work on the same dataset.

show abstract

Cyberbullying Detection with a Pronunciation Based Convolutional Neural Network

Cited by 91 publications

References 17 publications

Cyberbullying detection in social media text based on character‐level convolutional neural network with shortcuts

Cyberbullying detection in social media text based on character‐level convolutional neural network with shortcuts

Behavioral Analysis of User Data on Social Media Applications using Machine Learning Algorithms

Social Media Cyberbullying Detection using Machine Learning

Contact Info

Product

Resources

About