We are working on a manuscript for groundwater potential by considering different raster layers like Geomorphology, Geology, Soil texture, Elevation, Slope, Drainage density, Lineament density, Topographic Position Index, Topographic Wetness Index, Profile curvature, etc. Some of these parameters like Geomorphology, Geology, and Soil texture are qualitative variables while the others are quantitative variables. Hence, it is difficult to perform a collinearity test of these variables in their original form. So, I thanks in advance the experts in this field to help me in finding out the solution. My question is whether there is any specific technique to perform a collinearity test of taking both qualitative and quantitative variables? Whether the original data be resampled then a collinearity test be performed?
You can use Pearson's linear model for check multiple collineaarity in different variabels and also VIF test may be applyfor hidden correlation. Pearson's correlation is less than 0.7 and VIF value less than 10 is consider to be non correlated variables.
You can perform a colinearity test on quantitative data only via VIF calculation and if the value is more than 10 then one can drop that factor. Qualitative data is supposed to be orthogonal to each other and hence PCC and VIF not required on them .
- Badges
- Science topic
- Similar topics
- Linguistics
- Philology
- Classical Philology
- Manuscripts