Smart Detection and Prevention of Phishing Websites Using ML & NLP

Phishing attacks are one of the long-lasting problems in cybersecurity. Attackers are always changing their techniques to avoid being caught by conventional detection techniques. The techniques used currently are based on static blacklists or individual classifiers using features, but this does not completely represent the phishing attack. This project fills this critical information security gap by developing a hybrid real-time phishing detection system in the form of a Google Chrome extension. Our system uses a combination of two independent machine learning pipelines. In the first pipeline, we use thirty hand-crafted features to classify URL structure using a Gradient Boosting Classifier on 11,054 samples with 97.4% accuracy. In the second pipeline, we use raw HTML data from the webpage. We use a TF-IDF vectorizer on HTML data injected with semantic flag tokens. We use a Random Forest Classifier on 1,859 real site files with 82.80% accuracy. We combine the results using a weighted combination. We use a Retrieval Augmented Generation module, which uses a Google Gemini language model. We use a FAISS vector index to produce a risk explanation in a human-readable format. We use a trusted domain whitelist for false positives on legitimate banking and government websites. We use a pattern-based blocklist for piracy and malware domains. We use a popup interface to show users Safe or Phishing results. We use a full-screen alert overlay injected into the active tab for users. We use experimental results to show that our system performs better than any individual baseline

Phishing detection, Machine learning, Gradient Boosting Classifier, TF-IDF, Random Forest, Retrieval-Augmented Generation, Browser extension, URL feature extraction, HTML content analysis, Cybersecurity

1. L. Mat Rani, et al., “Feature Selection to Enhance Phishing Website Detection,” 2022. XGBoost, Random Forest, and Naïve Bayes models evaluated using URL-based datasets (79 and 56 features). DOI: 10.30880/jscdm.2023.04.01.003. [Google Scholar] [Crossref]

2. R.Alazaidah, et al., “Website Phishing Detection Using Machine Learning Techniques,” 2024. Includes RandomForest, J48, FilteredClassifier, and InfoGainAttributeEval for identifying optimal classifiers. UCI Repository datasets. [Google Scholar] [Crossref]

3. Alswailem, et al., “Detecting Phishing Websites Using ML,” 2019. Focus on feature selection, URL/page content, and browser extension integration using Random Forest. Dataset: PhishTank (12k phishing, 4k legitimate URLs). [Google Scholar] [Crossref]

4. V. Patil, et al., “Detection and Prevention of Phishing Websites Using ML,” IEEE, 2018. Multi-layer detection using blacklist/whitelist, heuristic rules, and visual similarity methods. Uses Alexa/WHOIS/Google datasets. (DOI Not Available). [Google Scholar] [Crossref]

5. F. Colhak, et al., “Phishing Website Detection Through Multi-Model,” 2024. Embedding fusion (tabular + NLP) with HTML-focused transformer models such as MLP, CANINE, RoBERTa. Dataset: ~65k samples from Aljofey’s dataset. [Google Scholar] [Crossref]

6. L. Thaci, et al., “Efficient Chrome Extension for Phishing Detection,” 2024. Chrome-extension-based detection using RF, SVM, and k-NN. Dataset: PhishTank, Alexa (11k pages). Accuracy: 95.6%. [Google Scholar] [Crossref]

7. E. Sri Vishva and D. Aju, “Phisher Fighter (URL & TF-IDF),” 2023. URL feature engineering and TF IDF analysis using DT, RF, SVM, and ANN. Dataset: Custom crawler-based browser dataset. [Google Scholar] [Crossref]

8. J. Kumar, et al., “Phishing Website Classification & Detection Using ML,” 2020. Lexical, URL, and page analysis using NB, RF, DT, LR, and KNN. Dataset: Custom balanced (1M URLs). [Google Scholar] [Crossref]

9. S. J. S., et al., “Detecting Phishing Website Using ML (Notification System),” 2020. Uses blacklist, pop-up admin notifications, and basic ML. Dataset: User-supplied blacklisted URLs. [Google Scholar] [Crossref]

10. D. Kalla, et al., “Phishing Website URLs Detection Using NLP/ML,” 2023. Ensemble classifiers including LR, KNN, RF, DT, NB, SVM, and LSVC. Dataset: Kaggle/multi-model dataset (150k fake, 360k legitimate URL. [Google Scholar] [Crossref]

• A Machine Learning Model for Predicting the Risk of Developing Diabetes - T2DM Using Real-World Data from Kilifi, Kenya
• AI-Powered Facial Recognition Attendance System Using Deep Learning and Computer Vision
• A Comprehensive Review on Brain Tumour Segmentation Using Deep Learning Approach
• A Scalable Retrieval-Augmented Generation Pipeline for Domain-Specific Knowledge Applications
• Predictive Maintenance in Semiconductor Manufacturing Using Machine Learning on Imbalanced Dataset