IOT SECURITY ENHANCEMENT VIA DECISION TREE RFE AND PEARSON CORRELATION

The recent growth on the Internet of Things (IoT) has been a major driver of the industrial revolution, especially with the rise of smart cities. IoT applications automate real-time processes, making life more efficient and convenient.These IoT-enabled systems enhance the quality of life, improve service quality, and increase operational efficiency.However, this technology also poses significant security risks.As smart devices handle sensitive data, there is a growing risk of cyber-attacks, data misuse, and privacy breaches, which can undermine user trust and hinder the progress of smart city development.To address these cybersecurity challenges, it's essential to develop a model that can protect IoT devices from various threats in real time.This paper proposes an innovative approach that combines efficient feature selection and fusion techniques to enhance intrusion detection in IoT systems.The proposed model starts by preprocessing the raw IoT data, preparing it for the next stages of analysis.The Decision Tree-based Pearson Correlation Recursive Feature Elimination (DT-PCRFE) method is then applied to select the most relevant features.By eliminating redundant and uncorrelated data, this step not only improves resource utilization but also reduces computational complexity, making the system faster and more efficient.After selecting key features, the model employs a feature fusion process, converting IoT device requests into word embeddings.This transformation strengthens the model's ability to detect diverse attack patterns, thus increasing its robustness.