REGRESSION ANALYSIS 



365 



2 5 10 20 30 



250 300 



00 no 120 liO 140 ISO 160 



DISTANCE (MILES) 



Figure 8.30. Standard deviation of height errors. 



meteorological parameters. The standard error as a function of target 

 position for 



eh = hiNs + a 

 is shown in figure 8.31; for 



eh = biNs + hiGo + a 

 in figure 8.32; and for 



eh = biNs + b2Go + 63 AA^ + a 



(8.49) 



(8.50) 



(8.51) 



in figure 8.33. 



The regression equations, (8.49) through (8.51), were used because the 

 respective figures demonstrate how each additional parameter enhances 

 the accuracy of the estimate. The parameters Go and AA'^ applied indi- 

 vidually, that is, 



and 



€h = bOo + a 



eh = bAN + a, 



were of significant value only for targets at low heights (h < 10,000 ft). 

 Examination of the figures shows that prediction of eh with A'^., provides 

 significant improvement over the mean for target heights above 15,000 ft. 

 The addition of Go improves the estimate for heights below 15,000 ft 

 and the addition of AA'^ provides a slight overall improvement. 



