uitSQLid: SQL Injection Detection Using Multi Deep Learning Models Approach

Injection attack is the most common risk in web applications. There are various types of injection attacks like LDAP injection, command injection, SQL injection, and file injection. Among these, SQL injection is often found in web applications that utilize databases. There has been significant research related to SQL injection attacks. In this study, we propose a system to detect SQL injection attacks by using various deep learning models (named uitSQLid). We employ natural language processing (NLP) approaches during the data processing stage and for feature vector creation. The three NLP models used are BERT, DistilBERT and Word2Vec. We utilize two deep learning models. The first one is Convolutional Neural Network model (CNN). The second one is Long Short-Term Memory networks model (LSTM). Experimental results show that our proposed system achieves an impressive accuracy of 99.73%. These results can be used to further develop related studies in the future.