Multiple linear logistic regression in Landslide analysis

[sanzida]

Hello everyone.

 

i want to prepare a susceptibility map of landslide, i have all the relevant maps in raster format, for instance landuse/landcover, slope, aspect, relative height, geology, drainage distance etc and the landslide inventory ( landslide record of my area as binary raster format).

 

now my question is how do i calculate the multiple linear regression  for each and every factors ( thematic layers) , so far i know its possible in spss software, but i dont know how to do this, i mean how could i incorporate my rasters and do this analysis,

 

can anyone give me the idea of this analysis, i know little bit of spss, so i need this help 

[3dbu]

here little thing about the topic:

http://webhelp.esri.com/arcgisdesktop/9.3/index.cfm?TopicName=Regression_analysis_basics

http://libraries.mit.edu/files/gis/regression.pdf

https://www.scdhec.gov/gis/presentations/ESRI_Conference_08/tws/workshops/tw_38.pdf

  may be usefull , regards 

[sanzida]

thanks for the link,,,,,,,,,,,,,,,but i want to know how to run this process, is this possible in arcgis 

[3dbu]

is possible, sure, but need have feature or raster data , or table data joined but make some analyst...hope be usefull..regards

[gisadept]

Hello sanzida,

 

If I want to do analyses of that kind, I rather avoid ArcGIS and alike. I like to have full control on the analysis. To say: in a statistical package you mostly are able to tune the calibration of estimation process.

You mentioned SPSS, I don't have any experience with; only with SAS. But in fact that doesn't matter.

 

The clue is to export your raster data to a ASCII file (e.g.: .txt). If all thematic maps have the same extent, and the same spatial resolution, you're of to start.

Do the export (please Google it, it's always interesting to learn by yourself) in only one program, so you are sure the export is done systematically in the same way.

Why is that important?

If the sequence of raster cells is equal, every n^thline/record/value in yhe ASCII file will correspond to exactly the same raster cell. So your data files are spatially coherent.

When the export is done, just import the ASCII files into your statistical software package (they are tables now, like 3dbu already mentioned) and there you go!

 

Good luck!