HYBRID DEEP LEARNING MODEL FOR MALWARE DETECTION IN ANDROID APPLICATIONS

Abstract

With the rapid growth of applications, accurately predicting future traffic patterns is crucial for efficient application management and optimization. This research presents a hybrid deep learning framework that combines CNN (Convolutional Neural Networks) and LSTM (Long Short-Term Memory) networks to enhance the accuracy of network traffic prediction. CNN is leveraged for feature extraction, capturing spatial dependencies within the data, while LSTM is employed to model temporal dependencies, enabling the framework to effectively learn long-term patterns in network traffic. Unlike traditional deep learning models, the proposed hybrid approach improves prediction accuracy by utilizing CNN's capability to extract essential features and LSTM's strength in handling sequential dependencies. The framework significantly outperforms existing deep learning models by addressing the challenges of feature extraction complexity and sequence learning, making it a robust solution for real-time network traffic forecasting. The Kaggle dataset is used to test the suggested model for malware prediction. The accuracy, precision, and recall of the suggested model are evaluated. According to analysis, the suggested model predicts malware with a 98 percent accuracy rate