--- a
+++ b/Semantic Features/StackEns.m
@@ -0,0 +1,20 @@
+function [ EnsemblePrediction, EnsembleError, EnsembleSuccess, beta ] = StackEns( outputs, labels, category )
+%StackEns Take the outputs from the various L0 classifiers and merge them
+%into a single L1 output. 
+%   Takes numerical values, outputs numerical values. Reccomend converting
+%   to classifications after this step. Uses linear regression over
+%   multiple variables (>2 variable) to merge them
+
+%Add with cross validation, currently training and testing on the same set
+
+trainFunc = @mvregress;
+evalFunc = @(X, trainFunc) X * trainFunc;
+
+[ EnsemblePrediction, EnsembleError, EnsembleSuccess, trainedStruct ] = CrossValLearn(outputs(:,:,category), labels(:,category), trainFunc, evalFunc);
+
+
+beta = reshape(cell2mat(trainedStruct), 20, []); %Convert to right format
+beta = mean(beta,2);
+
+%end
+