Threat Detection and Classification Using ML Techniques

The Research study detects intrusions from unwanted sources which could prove to be harmful for the users. These could be in the form of spams, ads, fake sites, etc. We could not always be on alert about them. This project then comes in action; it uses CICIDS2017 dataset for training machine learning models like, random forest, SVM and find which model gives best results with less time and more accuracy. This will predict whether certain traffic is dangerous and which is not dangerous. As the dataset is of 2017 it is a latest dataset and compatible with current traffic also, so this will create better predictions which will also be effectively applicable in current time.