|
Dear OF Team, we are in the process of analyzing our data derived via a plot-based systematic sampling grid in OF CALC. We would like to use the Error calculation script based on "Formulas for estimators and their variances in NFI 28.2.2014 K.T. Korhonen & Olli Salmensuu (Metla, Joensuu/Finland), point estimators" # @author Mino Togna, FAO However, this script does not produce meaningful results as it is, as far as I understand, adjusted to a Cluster-based approach. Am I correct? I there any way to adjust this accordingly? How would the script have to be changed for it to work in-line with our sampling design? Thank you for your support! Best regards, Alex |
|
Dear Alex, please check when your inventory case matches with this case, page 28 In order to use this particular error script (see p. 51 in the linked document), in Calc's Sampling design you should select CLUSTER sampling, and select APPLY CLUSTER SAMPLING ONLY FOR VARIANCE (button to be green). Then in Calc's Calculation screen, before running the scripts, select via Error button and QUANTITY (i.e. variable of interest) and category of interest. And then run the results into Saiku cubes. For other sampling strategies this particular error script does not work as such. We have unfortunately not programmed scripts for other sampling cases. If you want to apply a simple random sampling framework of estimators or any other method, then you can write all plot level results out from Calc into a CSV file. See linked document and scripts on p. 49-50. Then in Excel or R you can apply any method, or one option is also to use bootstrapping (sampling simulations) with the help of a handy Shiny tool in the SEPAL toolbox. Hope this helps you, Lauri |
