Enhancing Cybersecurity Through LSTM-Based Phishing URL Detection

Attackers over the years has frequently launch attacks on users of the internet in other to steal their personal vital information so as to achieve their fraudulent acts.To curb this attacks there is a need to develop a deep learning model that can conveniently detect URL that are phishing.Long short time memory (LSTM) model is used in this research, various approaches haven used but best result is yet to be preferred the approaches.LSTM is a variation of Recurrent Neural Network (RNN) Architecture designed to handle sequence related prediction problems and its does well when working on sequential data, such as speech recognition, natural language processing and Phishing detection.The two datasets from kaggle.com were used and the result shows that on Phishing_1 dataset (large dataset) has an accuracy of 0.8672 while on the second dataset Phishing_Legitimate_full (small dataset) has an accuracy of 0.9043 this therefore mean that LSTM can perform better on small datasets and there is tendency of result degradation on larger datasets.However, there are other metrics that makes LSTM a considerable model on larger datasets like the F1-score.