ABSTRACT A wireless sensor network (WSN) is essential for accurate localization in location‐based services and applications. It is difficult to scale and be accurate with geometric localization methods in complex environments. We propose machine learning models based on linear regression, random forests, and K‐nearest neighbors for detecting wireless signals. Received Signal Strength Indicator (RSSI) feature vectors are used to train a model, and supervised regression algorithms are used to predict node coordinates. An analysis of a large dataset is performed with KNN, whereas an analysis of a small dataset is conducted with linear regression. A decision tree's reliability is lowest because it has a high error rate and a wide range of errors. The study explains how users can select appropriate machine learning techniques to deploy WSNs efficiently by balancing model complexity, training data size, and localization accuracy.
Al‐Juboori et al. (2026) studied this question.