How does the RReliefF algorithm for regression work?
The original ReliefF algorithm for classification problems uses the concept of nearest hits and misses. I am confused how ReliefF can be used for regression?
You can use these two references:
This one explains the RReliefF Algorithm: "An Adaptation of Relief for Attribute Estimation in Regression, Robnik-Kononenko ,1995"
This one compares Relief, ReliefF and RReliefF: "Theoretical and Empirical Analysis of ReliefF and RReliefF, Robnik-Kononenko, 2003"
Taken from the reference 2003, page 28: "In regression problems the predicted value τ(·) is continuous, therefore (nearest) hits and misses cannot be used. To solve this difficulty, instead of requiring the exact knowledge of whether two instances belong to the same class or not, a kind of probability that the predicted values of two instances are different is introduced. This probability can be modelled with the relative distance between the predicted (class) values of two instances."