Method for Detecting Malicious Comments Using Deep Learning Models

This article presents a method for detecting malicious comments using a combination of deep learning models, specifically LSTM and the attention mechanism (Attention). This approach aims to improve the accuracy of identifying malicious comments by leveraging LSTM's sequential processing capabilities and the attention mechanism's focus on relevant elements. Technical details include the use of Word2Vec for word vector representation, LSTM for processing text sequences, and attention for highlighting important parts of comments. The impact of this method is an improvement in the detection of malicious comments, thereby contributing to online security.