predict error reduction

This code is for reproducing the results in “Active Learning with Interpretable Predictor”, Y. Taguchi, K. Kameyama, H. Hino, IJCNN2019.

setup

You have to Download dataset. We use 5 datasets here. wine quality, bike, CASP, housing, Freid

If you get these files, you put on the "dataset" folder.

experiment

If you done the setup, you can run the expriment.

Rscript exp_AL.R 1

c("loadCasp.R", "loadHousing.R", "loadredWine.R", "loadwhiteWine.R", "loadBike.R")

The args number represent dataset. The 1 is CASP, 2 is housing, 3 is redwine, 4 is whitewine and 5 is bike. When we run above command the result is expressed in "env" folder. You can run the plotmaker.R. Experimental result is in "env" folder