A Deep Learning-Based Classification Approach for Detecting Phishing Attack

In our research, we have to choose two models of machine learning for detecting phishing websites using the dataset. The data set was downloaded from the internet provider, specifically the Kaggle website. But before we go any further, there is a process to do before using any classification model, which is called pre-processing techniques. The following section explains in detail the dataset and the feature selections. The dataset has been downloaded from Kaggle supports a variety of dataset publication formats. Phishing attacks today are sophisticated and increasingly difficult to spot. A study conducted by Intel found that 97% of security experts fail at identifying phishing emails from genuine emails. The dataset includes 11430 URLs. The dataset is designed to be used as a benchmark for machine learning-based phishing detection systems. Features are from three different classes: 6 extracted from the structure and syntax of URLs, 24 extracted from the content of their corresponding pages, and 6 extracted by querying external services. The dataset is balanced; it contains exactly 50% phishing and 50% legitimate URLs.