Abstract : In this paper, we propose a novel logistic regression-based framework for classifying nodes in Internet of Things (IoT) networks as either trusted or blackhole nodes. The model leverages multiple behavioral features of nodes, including packet delivery ratio, packet loss, energy efficiency, responsiveness, cooperation, and reliability. These features are linearly combined using model coefficients and processed through a sigmoid function to yield a probability score between 0 and 1. Based on a predefined threshold, nodes are classified as trusted if the probability exceeds the threshold, and as blackhole nodes otherwise. A detailed mathematical example is provided to illustrate the implementation of the model. To evaluate its effectiveness, the proposed framework is tested under simulated IoT network conditions, and its performance is analyzed using key metrics such as packet delivery ratio, packet loss, end-to-end delay, detection accuracy, and routing overhead. Results demonstrate that the model efficiently identifies malicious nodes and enhances overall network reliability and security, offering a lightweight and effective solution for real-time threat detection in resource-constrained IoT environments.