Ensemble machine learning approach for classification of IoT devices in smart home

Abstract The emergence of the Internet Things (IoT) concept as a new direction technological development raises problems such valid and timely identification devices, security vulnerabilities that can be exploited for malicious activities, management devices. communication IoT devices generates traffic has specific features differences with respect to conventional This research seeks analyze possibilities applying classifying regardless their functionality or purpose. kind classification is necessary dynamic heterogeneous environment, smart home where number types grow daily. uses total 41 logistic regression method enhanced by supervised machine learning (logitboost) was used developing model. Multiclass model developed using 13 network generated Research shown it possible classify into four previously defined classes high performances accuracy (99.79%) based on flow Model performance measures precision, F-measure, True Positive Ratio, False Ratio Kappa coefficient all show results (0.997–0.999, 0.997–0.999, 0–0.001 0.9973, respectively). Such have its application foundation monitoring managing solutions large environments Industrial IoT, home, similar.