Question 568:

Asked on December 21, 2008

Tags: Trend Line , Linear Regression

1

Answer:

No answer provided yet.

Files Available For Download

1. Q568_1_trends.xls

1. See the attached excel file for the plot

2. We see a non-linear rise then dramatic fall then modest rise. One likely cause for the fluctuations in yearly plane shipments is due to changing demand in passenger demand. 1970-1972 were during recession years and could have caused the step decline in orders. Again a recession in 1981, which was followed by deregulation of the airlines and the shipments never recovered. One might also suspect a tipping point was reached in which the amount of air-travel leveled off-and with it the demand for new aircraft. Aircraft also likely became more reliable and also reduced demand.
3. The problem with a fitted trend is the data appear to be non-linear, that is no following a line either up or down and making it difficult to model with linear-regression.
4. See plot 2 and there appears to be another recessionary force at play here as we see the decline beginning in 2000-2002. Fitting a trend might be helpful as a crude estimate of; although it appears Plotting 2004 might show another non-linear relationship. This segment is better than the data from 1966 but not ideal.
5. I fit a linear trend see plot 2.
6. The earlier years appear to be sufficiently different in their makeup than the latter, that is the magnitude of planes is much greater than later making them a poorer predictor of more recent data. The regression equation is Planes = 182.21x + 666.84 where x is the year. For 2004 that would be year 12, so the prediction would be 182.21(12) +666.84 = 2853 planes in 2004.