Forecasting Demand of Indonesian Local Fashion Product Using Hybrid Method: K-Means Clustering and Decision Tree Algorithm

Abstract

Indonesian local fashion products are currently common in Indonesia because of highly level of competition where many business owner start their business in fashion local, higher demand from Indonesian citizen, efficient product development timeline and increasingly product diversity. The significant growth of Indonesian fashion local product has caused of a competitive business with the scope of Indonesian fashion local business. Accurate demand forecasting of such products become important in driving effective business especially in local fashion product and achieving a sustainable competitive advantage. In this study, forecasting demand of Indonesian local fashion product using hybrid method clustering K-means algorithm and classification method (Decision Tree) was conducted. Five independent variable/ attribute descriptive was conducted to classify the item using decision tree classification method. They are color, size, material, price, and motif. In this research, the forecasting was performed as mid-term forecasting method.