Efficient Predictive Analysis for Intrusion Detection Using Recurrent Neural Network Compared Over Convolutional Neural Network with Better Accuracy

This research focused on comparing the effectiveness of Recurrent Neural Networks (RNN) and Convolutional Neural Networks (CNN) for intrusion detection. Using a sample size of 10 for each network type, the study aimed to determine which algorithm offers better accuracy. The RNN, which uses multiple “trees” for classification-where each tree votes on the final decision-was tested for its accuracy in detecting intrusions. The analysis involved a pretest power calculation with 80 % power and a 95 % confidence interval, with 10 iterations for each method. The results showed that RNN achieved an accuracy of 97.46 %, while CNN achieved 93.12 %. Although the statistical test indicated no significant difference between the two methods (<tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$p=0.428; p&lt;0.05$</tex>), the RNN still demonstrated superior performance in terms of accuracy. Overall, the study concludes that the RNN is a more effective and faster algorithm for building machine learning models for intrusion detection compared to CNN.