Implementation of Data Mining Using K-Means Algorithm for Bicycle Sales Prediction

Abstract

During the pandemic, to reduce the number of Covid-19 spreads,  the government imposed social distancing and work from home (WFH) to reduce community activities outside the home. This caused people to have irregular patterns or lifestyles which less any physical activity . It surely can lower immunity system in which can increase the risk of being infected by the virus. Therefore, during the pandemic, sports or exercises become one of the activities that regularly carried out by the community to increase their immunity. One of the sports activities that can be done to maintain their immunity is cycling. Cycling itself is a light activity that can be practiced by all ages. This occasion is certainly a good marketing target for bicycle selling companies, but the company sometimes experiences problems regarding bicycle stocks that do not match with the consumer market target. The purpose of this study is to find out what types of bicycles are on demand by predicting bicycle sales and looking at the desired interests of the community. This study uses the K-Means Clustering algorithm. The results of the K-Means Clustering research are divided into three clusters; Cluster 1 with 209 members with the most interest in mountain bikes, Cluster 2 with 787 members with the most interest in folding bicycles, and Cluster 3 with 540 members with bicycle interests. Most of them are city bicycles, from the clustering process above, the Dunn Index validation (Dunn Index) can be obtained with a value of 0.1324532.