Phishing attacks through SMS messages have become a significant concern in the digital era, posing risks to individuals and organizations. This project aims to address this issue by investigating the application of Machine Learning (ML) and Natural Language Processing (NLP) techniques to identify phishing SMS messages.
The objective of this research was to develop a system capable of classifying SMS messages as either "smishing" (phishing SMS) or "not smishing" (legitimate SMS). By leveraging supervised ML classification models and NLP techniques, we aimed to accurately detect and mitigate phishing attempts.
We used a carefully curated dataset consisting of labeled SMS messages, categorized as either smishing or not smishing. The dataset was collected from various sources and included examples of phishing SMS messages commonly encountered in real-world scenarios.
1.Data Preprocessing: We performed data cleaning and preprocessing steps, including tokenization, lowercasing, and removal of stop words, special characters, and irrelevant information.
2.Feature Extraction: NLP techniques such as TF-IDF (Term Frequency-Inverse Document Frequency) and word embeddings (e.g., Word2Vec or GloVe) were applied to represent the SMS messages as numerical feature vectors.
3.Model Training: We employed several classification algorithms, including Random Forest, Logistic Regression, Support Vector Machines (SVM), and Naive Bayes, among others. These models were trained on the preprocessed SMS data with corresponding labels.
4.Model Evaluation: The trained models were evaluated using appropriate metrics such as accuracy, precision, recall, and F1-score. Additionally, techniques like cross-validation and hyperparameter tuning were employed to ensure robust performance.
5.Model Selection: Based on the comparative study of different classification algorithms, the Random Forest classifier demonstrated the highest prediction accuracy, achieving an impressive accuracy of 99.87%.
The research successfully developed a phishing SMS detection system using ML and NLP techniques. The Random Forest classifier emerged as the most effective model, providing a high prediction accuracy of 99.87%. This system has the potential to significantly enhance the detection and prevention of smishing attacks.
Phishing attacks through SMS messages pose a significant threat to individuals and organizations. This project demonstrated the feasibility of leveraging ML and NLP methods to effectively identify and classify phishing SMS messages. The developed system, with its high accuracy, can aid in mitigating the risks associated with smishing and contribute to strengthening cybersecurity in the digital era.
For more details, please refer to the project documentation and code available in the GitHub repository(https://github.com/suyogub11/Smishing-SMS-Phishing-Detection-using-Machine-Learning-and-NLP-Techniques/).