Abstract: Separating the short jobs from the long is a known technique to improve scheduling performance. This paper describes a method developed for accurately predicting the runtimes classes of the jobs to enable the separation. Our method uses the fact that the runtimes can be represented as a mixture of overlapping Gaussian distributions, in order to train a CART classifier to provide the prediction. The threshold that separates the short jobs from the long jobs is determined during the evaluation of the classifier to maximize prediction accuracy. The results indicate overall accuracy of 90% for the data set used in the study, with sensitivity and specificity both above 90%.

Anat Reiner-Benaim, Anna Grabarnick and Edi Shmueli, “Highly Accurate Prediction of Jobs Runtime Classes” International Journal of Advanced Research in Artificial Intelligence(IJARAI), 5(6), 2016. http://dx.doi.org/10.14569/IJARAI.2016.050606

@article{Reiner-Benaim2016,
title = {Highly Accurate Prediction of Jobs Runtime Classes},
journal = {International Journal of Advanced Research in Artificial Intelligence},
doi = {10.14569/IJARAI.2016.050606},
url = {http://dx.doi.org/10.14569/IJARAI.2016.050606},
year = {2016},
publisher = {The Science and Information Organization},
volume = {5},
number = {6},
author = {Anat Reiner-Benaim and Anna Grabarnick and Edi Shmueli}
}