ANALISIS CART (CLASSIFICATION AND REGRESSION TREES) UNTUK PREDIKSI PENGGUNA SEPEDA BERDASARKAN CUACA

Abstract

This study presents a rule-based classification model for user predictions based on weather. Cyclists are very popular because of the increased comfort and environment. The data used is public data from the Bike Sharing Dataset taken from Kaggle. The data has data on bicycle users every hour. With this data set, the authors managed to find the accuracy of the CART method which explains the accuracy of 96%. The results of this study show the estimated distribution under various bicycles with spatial variables, distribution of use and time including the most influential variables in the predictions of bicycle users.