Machine Learning Based Network Wide Ad Blocking System

In this modern world, internet has become a very basic requirement for every human being. This attracted Corporates to show advertisements on the internet. Meanwhile there were lots of scams, annoying advertisements, trackers, spy-wares, key-loggers, viruses and worms on the internet. There exists many software to block those malicious URLs but they work only on web browsers. These software's has predefined list of URLs which are matched and removed while loading a webpage causing a delay every time. A lots of new phishing URLs are created everyday which are not blocked by the software. To overcome all these problems an access point is created using raspberry pi in which the malicious URLs are blocked using DNS-Masq. A Decision-Tree based machine learning model is created to predict the maliciousness of URLs and added to blacklist regularly. So that people can escape from tracking, advertisements and malicious threats permanently and work without any fear and distractions.