Published Paper Details:

A potent technique for identifying Internet of Things (IoT) assaults and identifying novel forms of intrusion to get access to a more secure network is the application of deep learning in a variety of models. Where in there is a high False Positive Rate(FPR) in Network Intrusion Detection System(NIDS),with decent prediction rate. To reduce the FPR and give scalable solution, , we recommend using a Deep Learning (DL) model. By placing the model on the cloud, we can increase the NIDS's responsiveness during periods of high load, hence boosting availability. Because the model is installed as a micro service and is operating on Docker containers on the cloud instance, which can be accessed by REST APIs. According to the testing results, Long Short-Term Memory (LSTM) with two layers and Deep Neural Networks (DNN) with five hidden layers performed with a minimum accuracy of 88.75% and a maximum accuracy of 95.02%. The Random Forest technique in standard machine learning algorithms has achieved 86% accuracy. Network intrusions are becoming fraudulent and sophisticated. For any firm, having an appropriate network intrusion detection system nearby is the most important component.