| dc.description.abstract | Hiep Thanh Phat Company often forecasts demand by simple linear regression and the
company‟s forecasting is not accurate due to the high errors. Therefore, four methods
such as Multiple Linear Regression, Decision Tree, Random Forest and Artificial
Neural Networks will be used to solve this problem. Then Root Mean Square Error is
used to compare the above four methods. However, this thesis only focuses on 5
products. The business is suggested to choose the most suitable method in order to
predict demand correctly. Random Forest gives the lowest forecast errors for above 5
products. Then this study uses some methods to calculate the ordering policy.
Wagner-Whitin gives the most optimal results because it saves the most ordering cost
and holding cost | en_US |
