Securing Smart Homes: Machine Learning Approaches for Detecting Reconnaissance Attacks

The widespread use of smart home gadgets has resulted in heightened susceptibilities and the possibility of being exploited by cybercriminals, specifically through reconnaissance attacks that preceded more severe invasions. This study investigates the effectiveness of machine learning (ML) models in identifying recon attacks in smart home applications. The BoT-IoT dataset is used to assess four major ML algorithms: Random Forest (RF), Logistic Regression (LR), Extreme Gradient Boosting (XGBoost), and Light Gradient Boosting Machine (LightGBM). This dataset captures a wide range of network traffic events and attack kinds. The dataset undergoes preprocessing to increase the importance of features and decrease the number of dimensions, resulting in optimal performance of the model. Conducting thorough experimentation and cross-validation evaluates each model’s accuracy, precision, recall, and F1-score in detecting reconnaissance operations. The results suggest that ensemble models, specifically XGBoost and LightGBM, surpass conventional models in terms of accuracy and computing economy. This makes them reliable options for promptly detecting attacks in smart home situations. The findings highlight the capacity of sophisticated ML methods to improve the security and robustness of smart home networking against surveillance and subsequent attacks.