Iot-enabled Intrusion Detection in Cloud Using Machine Learning Architecture, Implementation and Evaluation
Year: 2026
Type: article
Abstract: This paper proposes a machine-learning-based intrusion detection framework for cloud-connected Internet of Things (IoT) environments.IoT deployments integrate heterogeneous devices such as RFID tags and readers, cameras, smart sensors, and smart meters.These devices continuously generate network and device-related records that may be transmitted to cloud infrastructure for centralized storage and processing.The resulting connectivity and scale increase the attack surface and create a need for an intrusion detection mechanism capable of distinguishing normal activity from malicious traffic.The proposed workflow collects IoT-related records, organizes them in a cloud data repository, and applies preprocessing using Min-Max normalization and Decimal Scaling.Hybrid feature extraction is performed using Discrete Wavelet Transform (DWT) and Principal Component Analysis (PCA) to represent relevant signal or traffic characteristics while reducing feature redundancy.For classification and optimization, the study considers a Multi-class Time Bound Random Forest Algorithm (MCTBRFA) integrated with Chaotic Spider Monkey Optimization (CSMO).The intended objective is to classify multiple categories of network activity and improve the selection of model parameters or features through the optimization stage.
Institution: Rathinam Technical Campus
Language: English
Subfield: Computer Networks and Communications
Field: Computer Science
Domain: Physical Sciences
Open Access status: bronze
OpenAlex ID: W7229602859
