If ! convergence don't predict test

src/BayesNet/BoostAODE.cc

@@ -32,23 +32,31 @@ namespace bayesnet {
     void BoostAODE::validationInit()
     {
         auto y_ = dataset.index({ -1, "..." });
-        auto fold = platform::StratifiedKFold(5, y_, 271);
+        if (convergence) {
-        dataset_ = torch::clone(dataset);
+            // Prepare train & validation sets from train data
-        // save input dataset
+            auto fold = platform::StratifiedKFold(5, y_, 271);
-        auto [train, test] = fold.getFold(0);
+            dataset_ = torch::clone(dataset);
-        auto train_t = torch::tensor(train);
+            // save input dataset
-        auto test_t = torch::tensor(test);
+            auto [train, test] = fold.getFold(0);
-        // Get train and validation sets
+            auto train_t = torch::tensor(train);
-        X_train = dataset.index({ torch::indexing::Slice(0, dataset.size(0) - 1), train_t });
+            auto test_t = torch::tensor(test);
-        y_train = dataset.index({ -1, train_t });
+            // Get train and validation sets
-        X_test = dataset.index({ torch::indexing::Slice(0, dataset.size(0) - 1), test_t });
+            X_train = dataset.index({ torch::indexing::Slice(0, dataset.size(0) - 1), train_t });
-        y_test = dataset.index({ -1, test_t });
+            y_train = dataset.index({ -1, train_t });
-        dataset = X_train;
+            X_test = dataset.index({ torch::indexing::Slice(0, dataset.size(0) - 1), test_t });
-        m = X_train.size(1);
+            y_test = dataset.index({ -1, test_t });
-        auto n_classes = states.at(className).size();
+            dataset = X_train;
-        metrics = Metrics(dataset, features, className, n_classes);
+            m = X_train.size(1);
-        // Build dataset with train data
+            auto n_classes = states.at(className).size();
-        buildDataset(y_train);
+            metrics = Metrics(dataset, features, className, n_classes);
+            // Build dataset with train data
+            buildDataset(y_train);
+        } else {
+            // Use all data to train
+            X_train = dataset.index({ torch::indexing::Slice(0, dataset.size(0) - 1), "..." });
+            y_train = y_;
+        }
+
     }
     void BoostAODE::trainModel(const torch::Tensor& weights)
     {
@@ -64,7 +72,7 @@ namespace bayesnet {
         double priorAccuracy = 0.0;
         double delta = 1.0;
         double threshold = 1e-4;
-        int tolerance = convergence ? 5 : INT_MAX; // number of times the accuracy can be lower than the threshold
+        int tolerance = 5; // number of times the accuracy can be lower than the threshold
         int count = 0; // number of times the accuracy is lower than the threshold
         fitted = true; // to enable predict
         // Step 0: Set the finish condition
@@ -115,15 +123,17 @@ namespace bayesnet {
             models.push_back(std::move(model));
             significanceModels.push_back(alpha_t);
             n_models++;
-            auto y_val_predict = predict(X_test);
+            if (convergence) {
-            double accuracy = (y_val_predict == y_test).sum().item<double>() / (double)y_test.size(0);
+                auto y_val_predict = predict(X_test);
-            if (priorAccuracy == 0) {
+                double accuracy = (y_val_predict == y_test).sum().item<double>() / (double)y_test.size(0);
-                priorAccuracy = accuracy;
+                if (priorAccuracy == 0) {
-            } else {
+                    priorAccuracy = accuracy;
-                delta = accuracy - priorAccuracy;
+                } else {
-            }
+                    delta = accuracy - priorAccuracy;
-            if (delta < threshold) {
+                }
-                count++;
+                if (delta < threshold) {
+                    count++;
+                }
             }
             exitCondition = n_models == maxModels && repeatSparent || epsilon_t > 0.5 || count > tolerance;
         }

src/BayesNet/Network.cc

@@ -132,9 +132,9 @@ namespace bayesnet {
     void Network::setStates(const map<string, vector<int>>& states)
     {
         // Set states to every Node in the network
-        for (int i = 0; i < features.size(); ++i) {
+        for_each(features.begin(), features.end(), [this, &states](const string& feature) {
-            nodes.at(features.at(i))->setNumStates(states.at(features[i]).size());
+            nodes.at(feature)->setNumStates(states.at(feature).size());
-        }
+            });
         classNumStates = nodes.at(className)->getNumStates();
     }
     // X comes in nxm, where n is the number of features and m the number of samples