Abstract :
This research discusses the usage of the train spare-parts using data mining techniques. The algorithm use is apriori algorithm by analyzing the high frequency pattern that is looking for a combination of items that fulfill the minimum of support and the minimum of confidence for establishment of association rules. The sample data are 13 items of train spare-parts in December 2016 from UPT. Balai Yasa Surabaya Gubeng. Results of multiplication support and confidence value of 67% with the accuracy of 42%. The result of applying the apriori algorithm can find the association or correlation of relation between a number of the train spare-parts that are often use simultaneously and can help the planning part of the UPT. Balai Yasa Surabaya Gubeng for the formulation or planning of the train spare-parts stock strategy.