using stratified splitobs in microtc

sadit · sadit · commit f3a04ac83add · 2025-04-08T09:14:37.000-06:00
diff --git a/src/utils.jl b/src/utils.jl
@@ -64,8 +64,7 @@ function microtc(
     )
 
     (traincorpus, trainlabels), (testcorpus, testlabels) = if sample < 1.0
-        obs = randobs((corpus, labels), ceil(Int, length(labels) * sample))
-        splitobs(obs; at)
+        splitobs(obs; at, shuffle=true, stratified=labels)
     else
         splitobs((corpus, labels); at, shuffle=true)
     end

Original file line number	Diff line number	Diff line change
`@@ -64,8 +64,7 @@ function microtc(`
`64`	`64`	`)`
`65`	`65`
`66`	`66`	`(traincorpus, trainlabels), (testcorpus, testlabels) = if sample < 1.0`
`67`		`- obs = randobs((corpus, labels), ceil(Int, length(labels) * sample))`
`68`		`- splitobs(obs; at)`
	`67`	`+ splitobs(obs; at, shuffle=true, stratified=labels)`
`69`	`68`	`else`
`70`	`69`	`splitobs((corpus, labels); at, shuffle=true)`
`71`	`70`	`end`