Uploaded estimated models

Author: lamferzon

Data/Models/PLS1_hat.mat

@@ -0,0 +1,46 @@
+clc
+clearvars
+warning off
+
+%% Data import
+
+data = readtable("..\Data\Sources\stell-faults.csv");
+faultClasses = data.Properties.VariableNames;
+faultClasses = faultClasses(28:end);
+rowLimits = [158 348 739 811 866 1268 size(data, 1)];
+colLimits = 28:size(data, 2);
+
+%% Models estimation and validation
+
+numIter = 50;
+testPercent = .30;
+
+% PLS2
+for nVar = 2:length(faultClasses)
+    subset = data(1:rowLimits(nVar), 1:colLimits(nVar));
+    X = subset{1:rowLimits(nVar), 1:28};
+    Y = subset{1:rowLimits(nVar), faultClasses(1:nVar)};
+    model = PLS(X, Y, "Algorithm", "PLS2", "Trace", "on");
+    model = model.orderAnalysis(numIter, "Trace", "on");
+    model = model.estimate(model.orderRed.bestAlpha);
+    model = model.predict;
+    model = model.validate(testPercent, "Repeatable", "true");
+    model = model.crossval("Repeatable", "true");
+    models_PLS2.("p" + nVar) = model;
+end
+save("..\Data\Models\PLS2_hat", "models_PLS2");
+
+% PLS1
+for nVar = 2:length(faultClasses)
+    subset = data(1:rowLimits(nVar), 1:colLimits(nVar));
+    X = subset{1:rowLimits(nVar), 1:28};
+    Y = subset{1:rowLimits(nVar), faultClasses(1:nVar)};
+    model = PLS(X, Y, "Algorithm", "PLS1", "Trace", "on");
+    model = model.orderAnalysis(numIter, "Trace", "on");
+    model = model.estimate(model.orderRed.bestAlpha);
+    model = model.predict;
+    model = model.validate(testPercent, "Repeatable", "true");
+    model = model.crossval("Repeatable", "true");
+    models_PLS1.("p" + nVar) = model;
+end
+save("..\Data\Models\PLS1_hat", "models_PLS1");