An Autoencoder-Enhanced Stacking Neural Network Model for Increasing the Performance of Intrusion Detection

Abstract

Security threats, among other intrusions affecting the availability, confidentiality and integrity of IT resources and services, are spreading fast and can cause serious harm to organizations. Intrusion detection has a key role in capturing intrusions. In particular, the application of machine learning methods in this area can enrich the intrusion detection efficiency. Various methods, such as pattern recognition from event logs, can be applied in intrusion detection. The main goal of our research is to present a possible intrusion detection approach using recent machine learning techniques. In this paper, we suggest and evaluate the usage of stacked ensembles consisting of neural network (SNN) and autoen-coder (AE) models augmented with a tree-structured Parzen estimator hyperparameter optimization approach for intrusion detection. The main contribution of our work is the application of advanced hyperparameter optimization and stacked ensembles together. We conducted several experiments to check the effectiveness of our approach. We used the NSL-KDD dataset, a common benchmark dataset in intrusion detection, to train our models. The comparative results demonstrate that our proposed models can compete with and, in some cases, outperform existing models.

RETRACTED ARTICLE: Fusion of transformer and ML-CNN-BiLSTM for network intrusion detection

“…The Auto-Encoder 1 for data enhancement and the Auto-Encoder 2 for feature enhancement use the same model structure, consisting of an input layer, a full connection layer, a batch regularization layer, and an output layer [27]. The specific model structure of AE is shown in Fig.…”

On the Effect of Intelligent Neurally Processed Features on COVID–19 Fake News Detection Performance

“…Gradient Boosting tried to reduce the errors of the weak learner models and combine the predictions from the weak learners to become a strong learner. The extension of Gradient Boosting technique is the XGBoost, which reduces the time computation and computational complexity of the Gradient Boosting technique [26,29,48].…”

“…False Negative -Negative samples incorrectly predicted as actual positives False Positive -Positive samples incorrectly predicted as actual negatives True Negative -The samples that are identified correctly as actual negatives Precision metric exhibits the accuracy of the positive class and measures whether the prediction of the positive class is correct as defined (26) and is given by, Precision = TP / (TP + FP) Recall is measured as the fraction of positive classes correctly detected to the total classes as defined in (27) and is given by, Recall = TP / (TP + FN) F1 Score is the weighted average score or harmonic mean of true positive (recall) and precision as defined in (28) and is given by, F1 Score = 2 * [(P * R) / (P + R)] Accuracy is measured as the fraction of total of True Positive, True Negative to the total of True Positive, False Positive, True Negative, and False Negative as defined in (29) and is expressed as,…”

Semantic Hashing for Fast Solar Magnetogram Retrieval

“…In order to perform this operation we used a fully-connected autoencoder (AE) to encode the previously obtained MRGD. The autoencoder are used in various machine learning tasks such as: image compression, dimensionality reduction, feature extraction, image reconstruction [4,5,6]. As autoencoders use unsupervised learning, they are perfect for generating semantic hashes.…”