Comment on how changing the model parameter values

Info iconThis preview shows page 1. Sign up to view the full content.

View Full Document Right Arrow Icon
This is the end of the preview. Sign up to access the rest of the document.

Unformatted text preview: ee Height 60 40 Tree Height 80 Tree Diameter 80 Tree Diameter 80 Tree Diameter 10 20 30 40 50 60 10 20 30 40 50 60 Lundqvist gamma2 original Lundqvist gamma2 20% smaller Lundqvist gamma2 20% greater 10 20 30 40 Tree Diameter 50 60 40 60 Tree Height 40 60 Tree Height 60 40 Tree Height 80 Tree Diameter 80 Tree Diameter 80 Tree Diameter 10 20 30 40 Tree Diameter 50 60 10 20 30 40 50 60 Tree Diameter 3. Comment on how changing the model parameter values changes the predictions. Is the change in response linear throughout the range of the prediction? Note that for linear models, when we change the intercept the line shifts upward or downward uniformly across the line, and when we change the slope the line remains a line but shows a greater or lesser incline (i.e. the line tilts at different angles, but remains a line). What happens to the lines during the changes induced by the sensitivity analysis in this example? It appears that for each model, changing the gamma0 value caused the most significant change in prediction of height, which makes sense since the gamma0 parameter was largest and thus would have the most variation. The lines do not fully retain their original shape when each gamma value is changed. Unlike for linear models, the change in response is not linear through the range of the prediction. This is clearly shown in the set of gamma2 values for the Gompertz model, where the dip at the start of the graph of predictions is steeper for the parameter value 20% smaller and more horizontal for the one 20% greater....
View Full Document

This note was uploaded on 09/09/2013 for the course BTRY 3020 taught by Professor Sullivan,p.j. during the Spring '09 term at Cornell University (Engineering School).

Ask a homework question - tutors are online