On Designing a Machine Learning Based Wireless Link Quality Classifier

TL;DR

This paper evaluates the design choices in developing a machine learning-based wireless link quality estimator, highlighting the importance of data balancing and feature engineering over the choice of ML algorithms.

Contribution

It provides a systematic analysis of how data resampling and feature engineering influence the performance of ML-based wireless link quality classifiers.

Findings

Resampling for class balance significantly improves performance.

Feature engineering has a larger impact than ML method selection.

Performance depends heavily on data preprocessing steps.

Abstract

Ensuring a reliable communication in wireless networks strictly depends on the effective estimation of the link quality, which is particularly challenging when propagation environment for radio signals significantly varies. In such environments, intelligent algorithms that can provide robust, resilient and adaptive links are being investigated to complement traditional algorithms in maintaining a reliable communication. In this respect, the data-driven link quality estimation (LQE) using machine learning (ML) algorithms is one of the most promising approaches. In this paper, we provide a quantitative evaluation of design decisions taken at each step involved in developing a ML based wireless LQE on a selected, publicly available dataset. Our study shows that, re-sampling to achieve training class balance and feature engineering have a larger impact on the final performance of the LQE than the selection of the ML method on the selected data.