IDS-IoT: Intrusion Detection System for the Internet of Things Using Enhanced Long-Short Term Memory

Gaurav Meena; Ajay Indian

doi:10.47852/AIAbonview52025066

Authors

Gaurav Meena Department of Computer Science, Central University of Rajasthan, India
Ajay Indian Department of Computer Science, Central University of Rajasthan, India

DOI:

https://doi.org/10.47852/AIAbonview52025066

Keywords:

security, IoT, intrusion detection system, LSTM, deep learning

Abstract

Network security and intrusion detection have become significant challenges with the emergent inclusion of Internet of Things (IoT) devices across several domains. In this article, we proposed an enhanced long-short term memory (E-LSTM) method for detecting intrusions in IoT-based datasets for the designing of more resilient and competent intrusion detection systems (IDSs) in the dynamic domain of IoT environments, as well as for the thoughtful selection of models appropriate for various dataset characteristics. Four distinct datasets were used in this study: KDD-Cup’99, NSL-KDD, UNSW-NB15, and CICIoT2023. The aim was to estimate and compare the performance across datasets. We provide subtle insights into model behaviors and their capacity to adjust to the particulars of each dataset through rigorous analysis. The proposed enhanced LSTM approach revealed significant differences in precision, recall, accuracy, and F1-score compared to other approaches like AdaBoost, DNN, RNN, and Logistic Regression. It was discovered that, for every dataset, the accuracy rate exceeded 95%.

Received: 22 December 2024 | Revised: 15 July 2025 | Accepted: 14 September 2025

Conflicts of Interest

The authors declare that they have no conflicts of interest to this work.

Data Availability Statement

The data that support the findings of this study are openly available in KDD-Cup’99 dataset at http://kdd.ics.uci.edu/databases/kddcup99/kddcup99.html, NSL-KDD dataset at http://www.unb.ca/cic/datasets/nsl.html, UNSW-NB15 dataset at http://doi.org/10.1109/MilCIS.2015.7348942, and CICIoT2023 dataset at https://doi.org/10.3390/s23135941.

Author Contribution Statement

Gaurav Meena: Conceptualization, Software, Validation, Formal analysis, Writing – original draft, Writing – review & editing, Supervision, Project administration. Ajay Indian: Methodology, Investigation, Resources, Data curation, Writing – original draft, Writing – review & editing, Visualization.