Link quality estimation based on over-sampling and weighted random forest

Aiming at the imbalance problem of wireless link samples, we propose quality estimation method which combines K-means synthetic minority over-sampling technique (K-means SMOTE) and weighted random forest. The adopts mean, variance asymmetry metrics physical layer parameters as parameters. is measured by level determined packet receiving rate. used to cluster samples. SMOTE employed synthesize samples for so make different levels reach balance. Based on forest, model constructed. In model, decision trees with worse classification performance are assigned smaller weight, better bigger weight. experimental results show that proposed has processed SMOTE. Furthermore, it than ones Naive Bayesian, Logistic Regression K-nearest Neighbour methods.