Question

A toy company buys large quantities of plastic pellets for use in the manufacturing of its products.  The production manager wants to develop a forecasting system for plastic pellet prices and is considering four different approaches and 6 different models.  He plans to use historical data to test the different models for accuracy.  The price per pound of plastic pellets (actual) has varied as shown:

Month	Price/Pound
1	$0.39
2	0.41
3	0.45
4	0.44
5	0.40
6	0.41
7	0.38
8	0.36
9	0.35
10	0.38
11	0.39
12	0.43
13	0.37
14	0.38
15	0.36
16	0.39

SIMPLE LINEAR REGRESSION

• Use all of the data, months 1-16, to calculate the regression equation for this data.  Use the months (1-16) as the independent variable (x) and price as the dependent variable (y).  Once you have the equation, forecast months 7-16 (enter 7, 8, etc. as the x value in the equation).  Calculate the MAD for this forecasting model.  Comment on the goodness of fit (R²) and significance of the model (F significance) to determine if this forecast model should be included in the consideration of the different approaches. (If the model is not significant, it cannot be considered, however, you must still make the forecasts and calculate the MAD)

Show your work in a table like the one below

SMA AP=3

SMA AP=4

WMA

ETC

Month

Actual

Forecast

Absolute Deviation

Forecast

Absolute Deviation

Forecast

Absolute Deviation

7

0.38

8

0.36

Answer 1

Data				Forecasts and Error Analysis
Period	Demand (y)	Period(x)		Forecast	Error	Absolute	Squared	Abs Pct Err
Period 1	0.39	1
Period 2	0.41	2
Period 3	0.45	3
Period 4	0.44	4
Period 5	0.4	5
Period 6	0.41	6
Period 7	0.38	7		0.397426	-0.01743	0.017426	0.000304	04.59%
Period 8	0.36	8		0.394559	-0.03456	0.034559	0.001194	09.60%
Period 9	0.35	9		0.391691	-0.04169	0.041691	0.001738	11.91%
Period 10	0.38	10		0.388824	-0.00882	0.008824	7.79E-05	02.32%
Period 11	0.39	11		0.385956	0.004044	0.004044	1.64E-05	01.04%
Period 12	0.43	12		0.383088	0.046912	0.046912	0.002201	10.91%
Period 13	0.37	13		0.380221	-0.01022	0.010221	0.000104	02.76%
Period 14	0.38	14		0.377353	0.002647	0.002647	7.01E-06	00.70%
Period 15	0.36	15		0.374485	-0.01449	0.014485	0.00021	04.02%
Period 16	0.39	16		0.371618	0.018382	0.018382	0.000338	04.71%
				Total	-0.05522	0.199191	0.00619	52.56%
Intercept	0.4175			Average	-0.00552	0.019919	0.000619	05.26%
Slope	-0.0028676				Bias	MAD	MSE	MAPE
						SE	0.026387
Forecast	0.36875	17
						Correlation	-0.47212
						Coefficient of determination	0.222896

This is a poor model because only 22% of the variation is explained.

This forecast model should not be included in the consideration of the different approaches.