Research on SQL Injection Detection Model Based on CNN

SQL injection attack has been listed as one of the top ten network application risks by the open web application security project (OWASP) for five consecutive years, and has always been the focus of network security research. In this paper, a SQL injection detection model based on Convolutional neural network (CNN) is proposed. The model is divided into model training stage and classification detection stage. The key parts are the word vector model and CNN model. Firstly, the SQL injection data set is pre-processed, word2vec, word vectorization and other steps to get the vectorized data, and then input to the CNN model for model training and classification detection in turn. The method proposed in this paper realizes the purpose of SQL injection detection, and provides a theoretical reference for researchers and database security maintainers of relevant departments in academic circles, which has a certain application value.