Research on SQL Injection Attacks using Word Embedding Techniques and Machine Learning
S. Venkatramulu, Sharfuddin Waseem Md, Arshiya Taneem, Sri Yashaswini Thoutam, Snigdha Apuri · Journal of Sensors, IoT & Health Sciences (JSIHS). · 2024
Most of the damage done by web application attacks comes from SQL injection attacks, in which the attacker(s) can change, remove, and read data from the database servers. All three tenets of security— confidentiality, integrity, and availability—are vulnerable to a SQL injection attack. Database management systems receive their queries in the form of SQL (structured query language). It is not a new field of study, but it is still important to detect and prevent SQL injection attacks. A method of SQL injection detection based on machine learning is proposed. Feature extraction, followed by implementing various word embedding techniques like count vectorizer, TFIDF vectorizer to process the text data which can effectively represent the SQLI features is performed. Classification algorithms like Logistic Regression, SVM and Ensemble techniques like XGBoost is employed. Our goal in doing this systematic review is to find a better machine learning model to detect SQL injection attacks via implementing different word embedding techniques. The accuracy and F1-score of machine learning algorithms in terms of predicting the SQLI query has been calculated and reported in this research paper.