Fake Account Detection in Twitter using Long Short Term Memory and Convolutional Neural Network
Louzar Oumaima, Ramdi Mariam, Baida Ouafae, Abdelouahid Lyhyaoui · International Journal of Engineering Trends and Technology · 2024
With the growing influence of social media platforms, the identification and prevention of fake accounts has become a crucial challenge for maintaining the integrity of online interactions. The proliferation of Online Social Network (OSN) platforms has given rise to a significant increase in the number of fake accounts, leading to numerous detrimental effects on online communities. Many strategies have been suggested by various communities to deal with false accounts in OSN. Therefore, this paper proposes an innovative approach for detecting fake accounts on Twitter based on the content of tweets. It incorporated Long Short-Term Memory (LSTM) and Convolutional Neural Network (CNN). It conducted this research in several processes, including data collection, data preprocessing, data reduction by applied Correlation-based Feature Selection (CFS) and Principal Component Analysis (PCA), and data classification. The suggested method, the LSTM-CNN approach, is to cluster more than 2,000,000 accounts from the MIB dataset, and the experimental results show that the approach has the highest accuracy of 98.95% compared with other research.