Stock Resupplying Strategy via Decision Tree Learning Algorithm

The swift development of economic advancement has caused demand for better lead-time management to sharply increase profit. Fuel stations require precision and reliable resupply timetable to avoid congested traffic, which may potentially slowing down transactions, which is why the author suggests using Decision Tree Learning method to deal with the problem. PT. Perkasa Makmur, a distributor of fuel in Sekadau, West Kalimantan, Indonesia, has never analyzed sales data to gain insights for a potentially better resupply timetable.
This research can be categorized as experimental qualitative research, with a population-spanning dataset collected within 36 months of 2021 to 2023 workyear. Sampling technique involves up to 720 data points using probability sampling. This study uses the last 5 months of data from 36 months of the collected dataset, starting in February 2024 and ending in June 2024. This research produces an optimal date for the fuel station to resupply its fuel stock within the weekly and monthly timeframes. The graph generated from leaf and branch nodes provides for the suggested date to resupply, with the accuracy rate of the method up to 76.5%.