Reliable Predictive Intervals for the Critical Frequency of the F2 Ionospheric Layer

摘要: This paper addresses the problem of reliably predicting an important HF communication systems parameter, critical frequency F2 ionospheric layer, with use a new machine learning technique, called Conformal Prediction (CP). CP accompanies predictions traditional algorithms measures confidence. The proposed approach is based on wellknown Ridge Regression but instead point produced by original method, it produces predictive intervals that satisfy given confidence level. Our experimental results extended dataset show obtained are well-calibrated and narrow enough to be useful in practice.