diff --git a/ccn2019-correct.R b/ccn2019-correct.R
index 14c4175..a07e4e5 100644
--- a/ccn2019-correct.R
+++ b/ccn2019-correct.R
@@ -36,52 +36,60 @@
 control <- trainControl(
   method = "repeatedcv",
   number = 5,
+  repeats = 2,
   classProbs = T,
   verboseIter = T,
-  summaryFunction = twoClassSummary,
   savePredictions = T,
-  sampling = "down"
+  sampling = "down",
+  selectionFunction = "oneSE"
 )
 
 pls.new_model <- train(
-  correct ~ t + l + s + v + n + tl + ll + sl + ul + vl,
+  correct ~ .-a-al-dp-cr-rt,
   data = train_data,
   method = "pls",
-  preProcess = c("center","scale"),
+  metric = "Accuracy",
+  tuneLength = 20,
+  preProcess = c("zv","center","scale"),
   trControl = control
 )
 
-pls.old_model <- train(
-  correct ~ t + n + v,
+plot(pls.new_model)
+plot(varImp(pls.new_model), main="Variables Importance for Correctness (New Model)")
+
+pls.common_model <- train(
+  correct ~ .-a-al-dp-cr-rt-tl-ul-sl-s-ll-vl-l,
   data = train_data,
   method = "pls",
-  preProcess = c("center","scale"),
+  metric = "Accuracy",
+  tuneLength = 20,
+  preProcess = c("zv","center","scale"),
   trControl = control
 )
 
-pls.new_model
-pls.old_model
-varImp(pls.new_model)
-varImp(pls.old_model)
+plot(pls.common_model)
+plot(varImp(pls.common_model), main="Variable Importance for Correctness (Common Model)")
 
-trellis.par.set(caretTheme())
-densityplot(pls.new_model, pch = "|")
-densityplot(pls.old_model, pch = "|")
 
-resamps <- resamples(list(new = pls.new_model, old = pls.old_model))
-summary(resamps)
-dotplot(resamps)
-diffValues <- diff(resamps)
-bwplot(diffValues, layout=c(1,3))
+#trellis.par.set(caretTheme())
+#densityplot(pls.new_model, pch = "|")
+#densityplot(pls.common_model, pch = "|")
+
+# Compile models and compare performance
+pls.models <- resamples(list(new = pls.new_model, common = pls.common_model))
+#DEBUG summary(pls.models)
+#DEBUG dotplot(pls.models)
+#DEBUG diffValues <- diff(resamps)
+bwplot(pls.models, metric = "Accuracy", layout=c(1,1), main="Correctness Model Performance")
 
 
 pls.new_predicted <- predict(pls.new_model, test_data, type="raw")
 pls.new_predicted_prob <- predict(pls.new_model, test_data, type="prob")
-pls.old_predicted <- predict(pls.old_model, test_data, type="raw")
-pls.old_predicted_prob <- predict(pls.old_model, test_data, type="prob")
+pls.common_predicted <- predict(pls.common_model, test_data, type="raw")
+pls.common_predicted_prob <- predict(pls.common_model, test_data, type="prob")
 
 confusionMatrix(pls.new_predicted, test_data$correct)
-confusionMatrix(pls.old_predicted, test_data$correct)
+confusionMatrix(pls.common_predicted, test_data$correct)
 
 library(pROC)
 par(pty="s")
@@ -97,7 +105,7 @@
     print.auc=T)
 
 plot.roc(test_data$correct,
-         pls.old_predicted_prob$CORRECT,
+         pls.common_predicted_prob$CORRECT,
          legacy.axes=T,
          lwd=4,
          col="darkgray",
@@ -107,8 +115,9 @@
          lty = 3,
          add=T)
 
-legend(100,100, legend=c("New Model", "Old Model"),
+legend(100,100, legend=c("New Model", "Common Model"),
        col=c("black", "darkgray"), lty=c(1,3),lwd=3, cex=0.8)
 
-# ggplot(pls.old_model, aes(d = pred$obs, m = pred$CORRECT)) +
-#   geom_roc()
+# requires plotROC package
+#DEBUG ggplot(pls.common_model, aes(d = pred$obs, m = pred$CORRECT)) +
+#DEBUG   geom_roc()

diff --git a/ccn2019-correct.R b/ccn2019-correct.R
index 14c4175..a07e4e5 100644
--- a/ccn2019-correct.R
+++ b/ccn2019-correct.R
@@ -36,52 +36,60 @@
 control <- trainControl(
   method = "repeatedcv",
   number = 5,
+  repeats = 2,
   classProbs = T,
   verboseIter = T,
-  summaryFunction = twoClassSummary,
   savePredictions = T,
-  sampling = "down"
+  sampling = "down",
+  selectionFunction = "oneSE"
 )
 
 pls.new_model <- train(
-  correct ~ t + l + s + v + n + tl + ll + sl + ul + vl,
+  correct ~ .-a-al-dp-cr-rt,
   data = train_data,
   method = "pls",
-  preProcess = c("center","scale"),
+  metric = "Accuracy",
+  tuneLength = 20,
+  preProcess = c("zv","center","scale"),
   trControl = control
 )
 
-pls.old_model <- train(
-  correct ~ t + n + v,
+plot(pls.new_model)
+plot(varImp(pls.new_model), main="Variables Importance for Correctness (New Model)")
+
+pls.common_model <- train(
+  correct ~ .-a-al-dp-cr-rt-tl-ul-sl-s-ll-vl-l,
   data = train_data,
   method = "pls",
-  preProcess = c("center","scale"),
+  metric = "Accuracy",
+  tuneLength = 20,
+  preProcess = c("zv","center","scale"),
   trControl = control
 )
 
-pls.new_model
-pls.old_model
-varImp(pls.new_model)
-varImp(pls.old_model)
+plot(pls.common_model)
+plot(varImp(pls.common_model), main="Variable Importance for Correctness (Common Model)")
 
-trellis.par.set(caretTheme())
-densityplot(pls.new_model, pch = "|")
-densityplot(pls.old_model, pch = "|")
 
-resamps <- resamples(list(new = pls.new_model, old = pls.old_model))
-summary(resamps)
-dotplot(resamps)
-diffValues <- diff(resamps)
-bwplot(diffValues, layout=c(1,3))
+#trellis.par.set(caretTheme())
+#densityplot(pls.new_model, pch = "|")
+#densityplot(pls.common_model, pch = "|")
+
+# Compile models and compare performance
+pls.models <- resamples(list(new = pls.new_model, common = pls.common_model))
+#DEBUG summary(pls.models)
+#DEBUG dotplot(pls.models)
+#DEBUG diffValues <- diff(resamps)
+bwplot(pls.models, metric = "Accuracy", layout=c(1,1), main="Correctness Model Performance")
 
 
 pls.new_predicted <- predict(pls.new_model, test_data, type="raw")
 pls.new_predicted_prob <- predict(pls.new_model, test_data, type="prob")
-pls.old_predicted <- predict(pls.old_model, test_data, type="raw")
-pls.old_predicted_prob <- predict(pls.old_model, test_data, type="prob")
+pls.common_predicted <- predict(pls.common_model, test_data, type="raw")
+pls.common_predicted_prob <- predict(pls.common_model, test_data, type="prob")
 
 confusionMatrix(pls.new_predicted, test_data$correct)
-confusionMatrix(pls.old_predicted, test_data$correct)
+confusionMatrix(pls.common_predicted, test_data$correct)
 
 library(pROC)
 par(pty="s")
@@ -97,7 +105,7 @@
     print.auc=T)
 
 plot.roc(test_data$correct,
-         pls.old_predicted_prob$CORRECT,
+         pls.common_predicted_prob$CORRECT,
          legacy.axes=T,
          lwd=4,
          col="darkgray",
@@ -107,8 +115,9 @@
          lty = 3,
          add=T)
 
-legend(100,100, legend=c("New Model", "Old Model"),
+legend(100,100, legend=c("New Model", "Common Model"),
        col=c("black", "darkgray"), lty=c(1,3),lwd=3, cex=0.8)
 
-# ggplot(pls.old_model, aes(d = pred$obs, m = pred$CORRECT)) +
-#   geom_roc()
+# requires plotROC package
+#DEBUG ggplot(pls.common_model, aes(d = pred$obs, m = pred$CORRECT)) +
+#DEBUG   geom_roc()
diff --git a/ccn2019-criterion.R b/ccn2019-criterion.R
index 3cdf184..66c66a0 100644
--- a/ccn2019-criterion.R
+++ b/ccn2019-criterion.R
@@ -57,19 +57,18 @@
   cr ~ t+l+s+n+tl+ll+sl+ul+vl,
   data = train_data,
   method = "pls",
-  preProcess = c("center","scale"),
+  preProcess = c("zv","center","scale"),
   trControl = control
 )
 
 pls.old_model <- train(
-  cr ~  t + n:v + n,
+  dp ~  n*t,
   data = train_data,
   method = "pls",
   preProcess = c("center","scale"),
   trControl = control
 )
 
-
 summary(pls.old_model)
 summary(pls.new_model)
 plot(varImp(pls.old_model))