PREDIKSI PENJUALAN PRODUK PAKAN TERNAK MENGGUNAKAN ALGORITMA C4.5 (STUDI KASUS: TOKO BERMITRA BERSAMA)

Abstract

Bermitra Bersama Store is a microenterprise that provides various types of livestock feed for various animals such as chickens, birds, fish, pigs and others. In order to determine what kind of products are marketable to consumers, research is needed to predict the sales of livestock feed products. The data used is derived from the sales of Bermitra Bersama Store over a 4-month period, from September to December, totaling 252 sales data. The method used in this research is the C4.5 algorithm, which is a classification algorithm of the decision tree type. The highest accuracy value was obtained from testing with 226 training data and 26 testing data, showing an accuracy rate of 81% with good classification accuracy. Additionally, a total of 14 samples were classified as 'marketable,' and 12 samples were classified as 'not marketable' during testing. The 'size' of the livestock feed was found to be the attribute that most influences the sales of livestock feed products, whether they are marketable or not.