Hybrid Top Features Extraction Model for Detecting X Rumor Events Using an Ensemble Method

The present research proposes a hybrid top feature extraction model for detecting rumor events from social media platforms by making use of an ensemble method that combines classifiers of Random Forest, LSTM, and XGBoost. The ultimate aim is to improve rumor detection accuracy by extracting and selecting the most contributive features from social media posts such as text, user handle, and topic information. In total, the dataset used comprises 62,445 entries: text, rumor labels, user handles, and related themes. The proposed model would thereby classify the posts as rumor or non-rumor with considerably greater accuracy by combining Random Forest for feature selection and XGBoost for ensemble learning with LSTM for sequential modeling of the text. The study established that the rumor detection accurately tackled the problem of classifying genuine against fake information using the combination of multiple models. From the experimental results, we note that the hybrid approach significantly outperforms the models in isolation and could serve as an important tool for real-time rumor detection on digital platforms.

   Keywords: Hybrid Model, Rumor Detection, Random Forest, LSTM, XGBoost, Feature Extraction, Social Media, Ensemble Learning, Text Classification, Machine Learning, Information Propagation