Correction: A. Alissa et al. Feature Subset Selection Hybrid Deep Belief Network Based Cybersecurity Intrusion Detection Model. Electronics 2022, 11, 3077

In the original publication [1], some equations and datasets are used without proper citation.

Concerns were raised on this matter, and the authors wish to make the following corrections:

In Section 2, a new paragraph is added after “datasets, respectively”.

Moreover, the author of [21] developed a hybrid DBN model that integrates Softmax Regression in the output layer and uses contrastive divergence to update hidden layers, aiming to enhance classification and intrusion detection in an ICS (industrial control system) environment using a dataset of 472,795 records. Although the model achieved an impressive 99.72% accuracy across seven attack types, the study has notable limitations: the dataset is proprietary and not publicly available, which restricts reproducibility; the author did not evaluate older DBN-based systems on the same dataset, making the claimed 5% improvement unverifiable; and the comparisons were limited to basic machine-learning methods rather than stronger deep-learning architectures. Despite these issues, the hybrid model still outperformed the tested baselines and demonstrated strong capability in handling large, complex ICS network data.

In Section 3.3, “as presented in [21]” is added after “Equations (4)–(6)”, “Form the equation” is changed to “From the Equation (7) as mentioned in [21]”, and “as mentioned in [21]” is added after “In Equation (8).”

In Section 4, “IDS dataset” is changed to “NSL-KDD dataset [27]”.

In Section 5, “developed” is changed to “evaluated”.

The following two references have been added to the references list:

21.

Süzen, A.A. Developing a multi-level intrusion detection system using hybrid-DBN. J. Ambient. Intell. Humaniz. Comput. 2021, 12, 1913–1923.

27.

Intrusion Detection on NSL-KDD. Available online: https://github.com/thinline72/nsl-kdd (accessed on 12 January 2022).

Due to the newly added references, subsequent references and the corresponding citations in the main text have been adjusted to align with the numerical order.

The authors clarify that the scientific conclusions remain unchanged. These corrections have been approved by the Academic Editor. The original publication has also been updated.