Securing the Web using URL based Analysis and Machine Learning Algorithms

As we have witnessed, the usage of the internet and smart phones has increased in recent years. This has caused a lot of fraud, scams, and the loss of personal information. Phishing attacks are the easiest way to get confidential information from users. This type of action is performed through phishing websites and various other platforms, including social media, e-commerce sites, etc. There are some websites that steal personal information, banking details, and important data like user IDs and passwords. All these are growing problems in the cyberworld that can affect the integrity of data, and they have to be resolved to achieve fairness in data. Phishing detection is a challenging problem, but many solutions were proposed, like rule-based analysis, list-based analysis, anomoly-based analysis, etc. So we have come up with a machine-learning-based URL analysis technique. This involves nine different algorithms and a dataset. The major algorithms will be gradient-boosting classifiers, multi-layer perceptrons, etc. We have implemented an interface that gives the output of the phishing probability of a given URL. The approach will depict that the proposed models have outstanding detection performance, accurate results, and a better success rate [1].