AI and IoT based Intrusion Detection System for Cybersecurity

A network's traffic can be monitored by an Intrusion Detection System (IDS), which then sends out warnings if any malicious behavior is detected. It's a program that checks if a website link leads to malicious content or if there was a policy violation on the network. As a case study for IDS, identifying phishing websites has been performed. One of the most effective tools for hackers to carry out a variety of security attacks is the widespread distribution of fake website Uniform Resource Locators (URLs) via messages, email, etc., with the goal of fooling unaware consumers into providing over sensitive information or worse. Traditional methods for detecting phishing websites, such as employing blacklists, are used to make predictions about phishing URLs. Traditional solutions based on blacklists are not effective at predicting many phishing websites, because so many new ones are created and released on the Internet every day. Three distinct Artificial Intelligence (AI) models-Random Forest (RF), Naive Bayes (NB), and XGBoost-are used to accurately detect dangerous URLs. Both the authentic and the fraudulent URLs were obtained from various sources. The AI model is trained with the retrieved relevant features. Following this, performance indicators are used to assess the quality of the AI models. According to the outcome of the testing data, the XGBoost model achieved a maximum accuracy of 96.98%, which is higher than the accuracy of the remaining models tested. Then the XGBoost model has been finalized and deployed as a website, anyone can immediately recognize the phishing URL.