| dc.description.abstract | Demand forecasting is a critical component in supply chain management, directly
influencing a company's profitability. Accurate demand forecasting helps firms increase its
profit by optimizing and distributing products at the right time and in the right quantities.
In this study, we focus on improving forecasting accuracy when historical data is limited
by applying machine learning models, specifically K-Nearest Neighbors (KNN) and KNN
with KD-tree algorithms. The models are trained and tested using actual data from a
mineral water factory. After that, we validate the models using metrics such as Root Mean
Square Error (RMSE), Mean Absolute Error (MAE), and training time. At the end, the
result show that combining KD-tree with KNN not only decreases computational time but
also maintains consistent forecasting performance. This method brings a notable advantage
in managing high-dimensional data and making efficient and accurate demand predictions. | en_US |
