Talk:Unsupervised classification
From AWF-Wiki
(Difference between revisions)
(Created page with "==Another statistical question== The map ''unsu_rej'' is supposed to contain the ''confidence level'' for each pixel. Does this mean that if one pixel has a value of 1.0000, it l...") |
|||
| Line 1: | Line 1: | ||
==Another statistical question== | ==Another statistical question== | ||
The map ''unsu_rej'' is supposed to contain the ''confidence level'' for each pixel. Does this mean that if one pixel has a value of 1.0000, it lies in a distance of max. 1 units from a mean calculated by i.cluster? Or does it mean something like „the pixel belongs to the assigned cluster with a propability of (100-1=) 99% ?“ [[User:Lburgr|- Levent]] 16:24, 15 January 2011 (CET) | The map ''unsu_rej'' is supposed to contain the ''confidence level'' for each pixel. Does this mean that if one pixel has a value of 1.0000, it lies in a distance of max. 1 units from a mean calculated by i.cluster? Or does it mean something like „the pixel belongs to the assigned cluster with a propability of (100-1=) 99% ?“ [[User:Lburgr|- Levent]] 16:24, 15 January 2011 (CET) | ||
| + | |||
| + | |||
| + | ==The Seeburg problem== | ||
| + | After applying the maximum likelihood classification we can already see that the little lake in the east called ''Seeburger see'' (information drawn from the nice plugin „Google layers“) disappeared. Watersheds are no longer represented by the map. Is this an acceptable loss of information or did something go wrong? [[User:Lburgr|- Levent]] 16:31, 15 January 2011 (CET) | ||
Latest revision as of 17:31, 15 January 2011
[edit] Another statistical question
The map unsu_rej is supposed to contain the confidence level for each pixel. Does this mean that if one pixel has a value of 1.0000, it lies in a distance of max. 1 units from a mean calculated by i.cluster? Or does it mean something like „the pixel belongs to the assigned cluster with a propability of (100-1=) 99% ?“ - Levent 16:24, 15 January 2011 (CET)
[edit] The Seeburg problem
After applying the maximum likelihood classification we can already see that the little lake in the east called Seeburger see (information drawn from the nice plugin „Google layers“) disappeared. Watersheds are no longer represented by the map. Is this an acceptable loss of information or did something go wrong? - Levent 16:31, 15 January 2011 (CET)
