Classification of Phishsing Websites

Justifications of selected approach, tools and techniques

I have compared different classification models while predicting phishing websites rather than limiting myself to only one model. This gives us an idea about which model is best suited for a perticular requirements.

For Ex.- Let's say got best accuracy in MLP but best Precision in Decision Tree which is an important comparison because we don't want a model which may declare a phishing website as legitimate because that may prove to be dangerous for a user. But also, we don't want a model whose accuracy is low since we get very less output. What we need is a good amount of accurate output with good precision also.

STEPS FOLLOWED:

1. Loading the dataset
2. Familiarizing with data
3. Visualizing data
4. Data Preprocessing and EDA
5. Hold Out validation
6. Model Building and Training (Decision Tree, Random Forest, MLP, XGB and SVM)
7. Comparison of Models