What is the exact procedure to make success and prediction rate curves in ArcGIS for GIS-based MCDA model of gully erosion susceptibility ?
I have made GIS-based MCDA model based on 11 gully erosion predisposing factors (Model values: MIN = 0, MAX = 4,969), and I would like to validate it with existing gullies data (n of gullies = 320), that I have obtained through OBIA approach. The gully validation layer is converted to binary format (gully = 1, no gully = 0).
Dear Fran
I have a complete (step by step) video tutorial about Susceptibility mapping and validation (success and prediction) using ArcGIS and Excel only. The comprehensive tutorials are on Udemy website.
- How to Produce Prediction Map in GIS With ArcGIS and Excel?, link: https://goo.gl/6c8A1H
- Prediction Maps & Validation using Logistic Regression & ROC, Link: https://goo.gl/y5EnLn
Happy learning!
Video guidence of Omar Althuwaynee are very helpful and easy to understand. Fran Domazetović i recommend you to use his method.
Hi,
An other easy way is to use the “area frequency” tool from ARCSDM toolbox for ARCGIS. Based on you gully points data and your susceptibility map, it will provide you a table with all the necessary outputs to built the success and prediction curves.
See : https://github.com/gtkfi/ArcSDM
Best regards,
Dear Fran,
Validation of your result is not a complex rocket science. The process is too easy, but you have to decide that which method do you want to use. The easiest method is index based method which can be easily done using ARCgis and Excel. For this method you have to reclassify your final result into 100 equal zones in gis and calculate cumulative area % of zones and reverse it (in excel), then calculate the no (or area) of your inventory (gully erosion) data and calculate cumulative% and make reverse in excel. Then use trapezoidal formula to calculate area under the curve.
And second popular method is receiver operating characteristic curve (ROC), for that you need to use statistical software (like spss, R etc), make gully pixel as 1 and absent pixel as 0. Then convert final data and inventory data as ascii and transfer in statistical software. where u can easily calculate AUCROC.
Remember, first method and second method is entirely different to each other.
If you have any confusion please inbox me.
Regards,
Thank you all for your answers, they were really helpful!
At the end I used ROC curves for my analysis, which I made with the help of Dinamica EGO softvare, which was really practical and easy to use.
(details can be found in: Mas, J. F., Soares Filho, B., Pontius, R. G., Farfán Gutiérrez, M., & Rodrigues, H. (2013). A suite of tools for ROC analysis of spatial models. ISPRS International Journal of Geo-Information, 2(3), 869-887.)
Hi, Mathieu Fressard
I am facing a problem with sdm tools. Could you help me to solve out?
My arcgis version is 10.1.
Regards
Dear Ananta,For success rate curve my final map is not classified in 100 classes.What happened ,i dont know.Can you give me stepwise solution.Thanku
- Badges
- Science topic
- Similar topics
- Biological Science
- Microbiology
- Virology
- Virus