|
Usually, woman likes to make up, and man, too. Razor is not a fashionable product, but it is a good gift for father day, since it is a daily supply.
Nowadays, most researches are focused on the sales prediction model for the product in technology, whereas the research on the sales prediction for the daily supply barely exists. The main purpose of this thesis is focused on the sales prediction on the razor. The goal for this research is to build suitable statistics model for sales prediction of razor.
Razor sales seem to have no seasonal trend. To get an accurate model for forecasting razor sales, flexible statistic model for sales forecasting are needed. At the moment since there are only 60 records available, exponential smoothing, ARIMA, X-11, transform function and Intervention model are used. The mean absolute percentage error (MAPE) are used to select the final suitable statistical model for forecasting the razor sales.
Using the MAPE as a criteria to choose a suitable model, the transfer function model is the best model. Moreover, from the performance level defined by Lewis (1982), our model locates at the good model range, 10% ~ 20%.
|