Hi,

Empirical risk minimization is simply the minimization of the loss on the training data given a set of hypotheses , e.g., logistic regression models.  To implement it in Logistic (in WEKA), you can set the ridge parameter to zero. (Note that the default ridge parameter is very small; thus, effectively Logistic applies empirical risk minimization by default.)

Minimizing the empirical risk doesn't make sense for unconstrained decision lists like those generated by PART because minimizing the loss on the training data will normally just overfit. Therefore, the PART classifier utilizes regularization in the form of pruning. FYI, PART implements a greedy heuristic method for finding a small as well as an accurate decision list.

HTH.

Samer, PhD.

Empirical risk minimization (ERM) is a principle in statistical learning theory which defines a family of learning algorithms and is used to give theoretical bounds on the performance of learning algorithms.See the related paper for

https://doi.org/10.1016/j.jksuci.2016.04.002