Improving detection accuracy of politically motivated cyber-hate using heterogeneous stacked ensemble (HSE) approach

Mullah, Nanlir Sallau; Zainon, Wan Mohd Nazmee Wan

doi:10.1007/s12652-022-03763-7

Cited by 3 publications

(3 citation statements)

References 36 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…According to Mullah & Zainon (2022) in their second paper that have been reviewed, the researchers built a novel heterogeneous stacked ensemble (HSE) classifier for detecting politically motivated cyber-hate on Twitter. We constructed a heterogeneous stacked ensemble with eight baseline estimators.…”

Section: Theme 1: Hate Speech Detectionmentioning

confidence: 99%

Hate Speech on Social Media in Malaysia - A Thematic Review

Azman¹,

Zamri²,

Hassim³

2023

IJARBSS

View full text Add to dashboard Cite

Background -Hate speech is a problem in Malaysia, with various groups being targeted based on their ethnicity, religion, and political views. In recent years, there have been cases of individuals being charged with hate speech for comments made on social media or during public speeches. The government has also taken steps to combat hate speech by promoting interfaith and interethnic harmony through various initiatives and programs. GAP -Despite the increasing number of research on Hate Speech, there is lack of studies on thematic review hate speech in social media. Aim of The Research -Therefore, this article will review the current literature on hate speech, reporting of hate speech trend research on hate speech on social media especially in Malaysia. Methodology -A list of keywords related to hate speech followed by a filter using inclusion criteria from SCOPUS, RESEARCH GATE and WoS databases identified 53 peer-reviewed journal articles from 2018-2023. A thematic review have been made by the topic of hate speech literature using Atlas-ti 23. Findings -As a result, out of 53 of the papers, only 20 of research and studies are being reviewed. This also conclude a word cloud and three main theme of those papers are revealed, which is 35% from the data is by theme Hate Speech Detection, 55% is literature from Hate Speech Issues and 10% are Hate Speech in Covid-19 Pandemic. Contribution -The paper contributes to analysing the trends of hate speech in social media especially in Malaysia. Future Studies / Benefits Of Research -Researcher hope to give a great input to next scholars in the studies of hate speech on social media.

show abstract

Section: Theme 1: Hate Speech Detectionmentioning

confidence: 99%

Hate Speech on Social Media in Malaysia - A Thematic Review

Azman¹,

Zamri²,

Hassim³

2023

IJARBSS

View full text Add to dashboard Cite

show abstract

“…The weaker learners employed are ensemble classifiers or boosted ensemble algorithms such as GBC, RF, XGB, DT, and ABC. Boosting the ensemble is a learning strategy that corrects the prediction errors made by previous models to incrementally enhance its accuracy [15]. The boosting ensemble is used as the weaker learner, and stacked generalization is employed to build the primary classifier.…”

Section: Boosting Ensemblesmentioning

confidence: 99%

“…F1-score and Matthews correlation coefficient (MCC) is the recommended evaluation metrics for the data set with an imbalanced class distribution [15]. The confusion matrix (CFM) gives the model's performance numerical values.…”

Section: Model Evaluation Metricsmentioning

confidence: 99%

A Framework for Optimizing Ensemble Learning for Offensive Tweet Identification

Mullah

Zainon

Wahab

2022

Preprint

Self Cite

View full text Add to dashboard Cite

Online derogatory comments are ubiquitous on social media and areraising serious concerns across the globe. Social media data is riddenwith high dimensional search space due to noise, redundant features,and non-standardized writing style. These problems lead to high computationalcosts, longer training time, and low predictive accuracy inmachine learning models. The researchers proposed a framework for optimizingstacked generalized ensemble learning to address these problemsand enhance model performance. The main components of the frameworkinclude feature optimizer (FO), ensemble classifiers, and stratifiedK-foldCV (skfCV) through stacked generalization ensemble architecture.The ensemble classifiers, FO, and skfCV components make our methodstable and computationally efficient with the best performance. Theproposed method was validated using three benchmark datasets. The proposed method outperformed the state-of-the-art results in all theevaluation metrics used in those three articles adopted for comparison.

show abstract

English Long and Short Sentence Translation and Recognition Method Based on Deep GLR Model

Wang

Zhao

2022

Computational Intelligence and Neuroscience

View full text Add to dashboard Cite

The translation recognition of English long and short sentence information is an important issue to obtain the focus and core of English articles. Based on the deep GLR model, this paper constructs a method framework for English long and short sentence translation and recognition, using different embedding layer parameter initialization methods and using multi-layer computing methods in the sentence decoder. The initial corpus text is segmented and tagged with part-of-speech, then, the part-of-speech tag is appropriately corrected to reduce ambiguity, and then it is manually syntactically tagged. In the simulation process, the English long and short sentence summary and translation components are designed and developed, which can accurately and efficiently obtain the key information of English long and short sentences. The experimental results show that the English long and short sentence translation and recognition method of the deep GLR model improves the accuracy of the model parameters. In terms of model structure, the deep GLR value can be improved by 70.77% by reproducing the multi-layer representation fusion of semantic translation; in terms of data enhancement, the deep GLR value can be increased by 70.35% by means of “back translation,” and the improved model is effective. It promotes the translation and recognition generalization ability of English long and short sentences.

show abstract

Improving detection accuracy of politically motivated cyber-hate using heterogeneous stacked ensemble (HSE) approach

Cited by 3 publications

References 36 publications

Hate Speech on Social Media in Malaysia - A Thematic Review

Hate Speech on Social Media in Malaysia - A Thematic Review

A Framework for Optimizing Ensemble Learning for Offensive Tweet Identification

English Long and Short Sentence Translation and Recognition Method Based on Deep GLR Model

Contact Info

Product

Resources

About