Department of Industrial Engineering2024-11-0920079780-4445-3157-51570-794610.1016/S1570-7946(07)80187-82-s2.0-43049159451http://dx.doi.org/10.1016/S1570-7946(07)80187-8https://hdl.handle.net/20.500.14288/14741The early prediction of activity related characteristics of drug candidates is an important problem in drug design. The activities of drug candidates are classified as low or high depending on their IC50 values. Since experimental determination of IC50 values for a vast number of molecules is both time consuming and expensive, computational approaches are employed. In this paper, we present a novel approach to classify the activities of drug molecules. We use hyper-boxes classification method in combination with partial least squares regression to determine the most relevant molecular descriptors of the drug molecules in efficient classification. The effectiveness of the approach is illustrated on DHP derivatives. The results indicate that the proposed approach outperforms the other approaches reported in the literature.Industrial engineeringJournal Articlehttps://www.scopus.com/inward/record.uri?eid=2-s2.0-43049159451anddoi=10.1016%2fS1570-7946%2807%2980187-8andpartnerID=40andmd5=9ff19a9ccb74b3fd035e83067ea957e5287727400165N/A8645