The ever-changing nature of wireless communication environments presents significant challenges in maintaining consistent communication quality, particularly in Wi-Fi networks. Predicting Wi-Fi link quality is essential for optimizing network performance, ensuring reliable connectivity, and minimizing service disruptions. This study explores the application of machine learning models for accurate Wi-Fi link quality prediction, focusing on low-complexity models that can be deployed in resource-constrained hardware environments. We evaluate several machines learning algorithms, including Random Forest, Stacking Classifier, Voting Classifier, and Decision Tree, which leverage a linear combination of exponential moving averages for performance forecasting. The models are trained on a dataset derived from real-world Wi-Fi link data, and the output is categorized into three levels of link quality: Very Good, Good, and Poor. Experimental results demonstrate the potential of these models in providing reliable, efficient, and scalable predictions, even in environments with limited computational resources.

Keywords: Wi-Fi link quality, machine learning, Random Forest, Stacking Classifier, Voting Classifier, Decision Tree, exponential moving averages, low-complexity models, link quality prediction, wireless communication.