Question

Home Depot is trying to determine which forecasting method works best to predict the number of gardenias to order from Bonnie’s Plant Farm. Below is a listing of the number of gardenias sold each day last week and the predictions for two different methods.

Day Actual Number of Gardenias Predicted Demand Trend-Seasonal Linear Regression

Monday 50 20 80

Tuesday 80 52 100

Wednesday 60 45   120

Thursday 52 90 140

Friday 125 197 160

Saturday 250 580 180

Sunday 175 390 200

A) Compute the MAD and Bias for each technique.

B) Construct a sentence using the MAD and Bias that tells the average person about each forecasting technique.

C) Which forecasting technique, Trend Seasonal or Linear Regression, gives the best results? How did you derive this conclusion?

Answer 1

Solution

A) Refer below for MAD and Bias calculation and its values

B) MAD is used for calculation of demand variability and Bias is general direction or tendency of error.

In this case, more Bias exists if we use Trend Seasonal method than Linear Regression. It also means cumulative forecast of Trend Seasonal method is having more negative tendency than Linear regression. Similarly, MAD by using Linear regression is lesser than Trend Seasonal method hence LR method is better.

C) Based upon above calculations, Linear Regression looks better than Trend Seasonal prediction based on values of MAD and Bias.