TY  - JOUR
AU  - Harit, Vibhor 
AU  - Dahiya, Rajeev 
AU  - Garg, Umang 
AU  - Kumar, Anil 
PY  - 2026
TI  - An Explainable AI (XAI) Framework for Intrusion Detection and Identification to Secure IoT Networks
JF  - Journal of Computer Science
VL  - 22
IS  - 7
DO  - 10.3844/jcssp.2026.2274.2290
UR  - https://thescipub.com/abstract/jcssp.2026.2274.2290
AB  - The Internet of Things (IoT) encompasses a wide range of applications from wearable devices to smart homes, as well as industrial automation and smart cities. As IoT networks and devices continue to grow rapidly, security has become crucial. There are many techniques utilized to provide security aspects in IoT networks. These security aspects focus on mitigating risks and fixing vulnerabilities related to connected devices. Due to the sheer quantity, complexity, and amount of data, the increasing number of IoT devices does in fact increase the risk of new attacks and vulnerabilities. An intrusion detection system is supposed to be powerful for the detection and identification of attacks. Some intelligent models are trained with a set of features and evaluated by applying a filter-based approach for selecting significant features. Deep learning models have intricate architectures with many layers and neurons; they are sometimes seen as opaque and challenging to understand. Explainable Artificial Intelligence (XAI) methods have been devised to provide information about models. The utilization of XAI techniques, including Local Interpretable Model-agnostic Explanations (LIME) and Shapley Additive Explanations (SHAP), helps better understand the rationale behind model predictions and encourages the user to understand the model behavior better.