Skripsi
DETEKSI SERANGAN SQL INJECTION MENGGUNAKAN DEEP NEURAL NETWORKS
SQL Injection has been classified by OWASP (Open Web Application Security Project) as one of the most damaging attacks in the past 15 years. This research aims to develop an IDS (Intrusion Detection System) capable of detecting SQL Injection attacks through HTTP POST Requests. The study employs a neural network-based computational model known as DNN (Deep Neural Networks), using tabular data that undergoes pre-processing before being utilized to build the DNN model. This data consists of statements labeled to indicate whether they are SQL Injection attacks or not. Subsequently, the model will be configured and trained using different parameters. The best-performing model will be integrated with a packet capturer, forming an IDS that will be tested against real attacks. The research findings indicate that the best-performing model, configured with the parameters ngram_range (1, 2), min_df 4, max_df 0.8, and trained for 5 epochs, achieves an accuracy of 96.0% on test data and a loss of 20.0%. Furthermore, testing the integrated IDS with the best model against real attacks showed a confusion matrix with values of 2935 (True Negatives), 841 (True Positives), 137 (False Negatives), and 286 (False Positives).
Inventory Code | Barcode | Call Number | Location | Status |
---|---|---|---|---|
2407003981 | T149261 | T1492612024 | Central Library (REFERENCES) | Available but not for loan - Not for Loan |
No other version available