Abstract:

Detection of anomaly and attack identification is some of the major concern in IoT domain in recent days. With the exponential use of IoT based infrastructure in every domain, threats and anomalies are amplifying adequately. Attacks such as malicious operations, spying, service denial etc. are the main cause for failure in IoT system. So, developing an efficient model to identify and decipher such complex problem is always been a challenging task. Rather some of the machine learning based models is developed to solve such problem, but due to highly nonlinear nature of the data, such methods seem to be failed to prove the efficacy. With the combination of several models, ensemble learning helps to enhance the performance of machine learning methods. As compared to any single method, the ensemble learning based models are highly predictable for large dimensional data. In this paper, an adaptive boosting based model has been proposed to identify the anomaly in IoT based environment. The performance of the proposed method is compared with several other competitive machine learning based methods and found to be superior with all the considered metrics.

Keywords: