Forum: open-discussion
Monitor Forum | 
Start New Thread
| RE: costdistances less than transition costs [ Reply ] By: Jacob van Etten on 2010-10-20 20:37 | [forum:3529] |
|
Yes and yes! The vignette explains why this is so, but the short explanation is that this setup allows us to use sparse matrices. |
|
| RE: costdistances less than transition costs [ Reply ] By: Tabitha Graves on 2010-10-20 16:38 | [forum:3525] |
|
Got it-- one last question then (er 2). The output matrix is in terms of resistance = least cost then, right? And resistancedistance() is also reported in terms of the effective resistances as the command sounds like? Thanks again! |
|
| RE: costdistances less than transition costs [ Reply ] By: Jacob van Etten on 2010-10-20 15:26 | [forum:3524] |
| costDistance converts the matrix into a graph and assigns weights of resistance = 1/conductance to the edges. Then it calculates minimum resistance paths using this graph. | |
| RE: costdistances less than transition costs [ Reply ] By: Tabitha Graves on 2010-10-20 15:11 | [forum:3523] |
|
So in that case, high cell costs lead to low values for conductance. So does costdistance actually find the maximum conductance path then? The conductance approach makes a lot of sense for the circuit calculations. I'm just trying to wrap my poor little brain around how it works for least cost path. Thanks for taking the time to explain! |
|
| RE: costdistances less than transition costs [ Reply ] By: Jacob van Etten on 2010-10-19 21:14 | [forum:3520] |
| Yes, or the formula entered in the TransitionFromRaster function should at least convert them to proper conductance values. | |
| RE: costdistances less than transition costs [ Reply ] By: Tabitha Graves on 2010-10-19 16:59 | [forum:3517] |
|
Ok, so the values in the original raster brought in also need to reflect conductances, rather than costs or resistances then right? I misunderstood the vignette and thought that it would take costs and calculate conductances when it made the transition matrix. Thanks for clarifying! |
|
| RE: costdistances less than transition costs [ Reply ] By: Jacob van Etten on 2010-10-19 08:13 | [forum:3515] |
|
Dear Tabitha, The values in the transition matrix are supposed to be conductance values. The function will then calculate the friction as 1/conductance. The vignette explains this. |
|
| costdistances less than transition costs [ Reply ] By: Tabitha Graves on 2010-10-18 22:15 | [forum:3514] |
|
I was examining the values in the resulting costdistance matrix and the highest value is a little over 4 in my dataset, while the lowest transition cost is actually 270. Is there some sort of standardization that is occuring? Thanks. -Tab |
|
