PENINGKATAN AKURASI KLASIFIKASI INTERAKSI FARMAKODINAMIK OBAT BERBASIS SELEKSI PASANGAN OBAT TAKBERINTERAKSI

Hilma Mutiara Winata (Department of Statistics, IPB University, Indonesia)
Farit Mochamad Afendi (Department of Statistics, IPB University, Indonesia)
Anwar Fitrianto (Department of Statistics, IPB University, Indonesia)

Indonesian Journal of Statistics and Its Applications
Vol 3 No 3 (2019)

Identifying the pharmacodynamics drug-drug interaction (PD DDI) is needed since it can cause side effects to patients. There are two measurements of drug interaction performance, namely the golden standard positive (GSP) which is the drug pairs that interact pharmacodynamics and golden standard negative (GSN), which is a drug pairs that do not interact. The selection of GSN in the previous which studies were only selected randomly from a list of drug pairs that do not interact. The random selection is feared to contain drug pairs that actually interact but have not been recorded. Therefore, in this study the determination of GSN was carried out by, first, grouping drug pairs included in the GSP using the DP-Clus algorithm with certain values of density and cluster properties. Then the drugs in different group would be paired and only the drug pairs in the GSN list are selected. It was found that our new proposed classification method increases the AUC value compared to the results obtained by random selection of GSN.