Create a discrimination model

1

I want to create a discrimination model to classify honey and fortified honey, and use it in a set of new spectral data. In the stage of creating the model, of discrimination that I am doing by FDA (Factorial Discrimant Analysis), an error appears and the model is not created. Would you know how to tell me the reason for the error?
The parameters I use to create the model are:
“Discrimination”
“FDA” Factorial Discrimant Analysis
Select X data: xCAL(CovSel on (PCA on SNV(xCAL.tab):scoresXyCAL.tab(c2)):Xdata(30varsel))
Select Y data: yCAL(yCAL.tab)
Column of y class chosen for the calculation: c2: V1
Number of blocs for cross-validation: 10
Number of discriminant variables(DV): 10
Distance choice: Mahalanobis
Export confusion tables (CAL) for each discriminant variables (DV) models: no

Stoud: Free Access Chemometric Toolbox (FACT) version 0.10
getting started at: http://atoms.scilab.org/toolboxes/FACT
Load macros
Load gateways

!–error 4
Undefined variable: test
at line 46 of function conj2dis called by :
at line 31 of function cvclass called by :
at line 9 of function fda called by :
res_da=fda(x,y.d(:,ycol),10,10,0);

                        !--error 4

Undefined variable: res_da

                        !--error 4

Undefined variable: res_da

                                                             !--error 248

Wrong value for argument #2: Valid variable name expected.

Stderr : Wrong value for argument #2: Valid variable name expected.

The URL is https://vm-chemflow-francegrille.eu/

2

I am not familiar with this tool but if the above are inputs then you might want to first double check the syntax. The error describes Wrong value for argument #2: Valid variable name expected. which tells me that you might have provided a bad variable name.

You should also look into the tools arguments and figure out which argument is #2. The administrator of the galaxy instance you are using should be able to help you with that.

1 Like
3

I had chosen the wrong number of variables, now I managed to solve this problem. @innovate-invent Thank you very much for answering!