IoT models are becoming more advanced as demand and growth in the Internet of Things (IoT) automated network system grows. People are becoming accustomed to data-driven infrastructure, and this is leading the research more on Machine Learning based applications alongside IoT. Human Life, at present, uses IoT and Machine Learning techniques everywhere. In medicine interpretation of ECG, pattern finding in genomic data, brain signal modelling all these complex tasks require the introduction of machine learning approaches. IoT devices use a wireless channel to transmit data which makes them a vulnerable target. A typical communication attack on a local network is restricted to local nodes or small local domains, however an attack on an IoT system spans a larger geographic area and has major repercussions on the IoT sites. The primary goal of our system is to develop a smart, secured and reliable IoT based infrastructure that can detect its vulnerability along with having a secure firewall against all cyber-attacks which allows it to recover itself automatically. Here, a Machine Learning based solution will be proposed which can detect and protect the system when it is in an abnormal state.