Merge pull request #16 from sdevenes/feature/paper

Feature/paper

scripts/database.py

@@ -113,4 +113,4 @@ def get(protocol, subset, classes=CLASSES, variables=VARIABLES, setname='csh101'
     retval = split_data(load(setname), subset, PROTOCOLS[protocol])
     varindex = [VARIABLES.index(k) for k in variables]
     retval = dict([(k, retval[k][:, varindex]) for k in classes])
-    return np.array([retval[k] for k in classes])
+    return np.array([retval[k] for k in classes], dtype=object)

scripts/paper.py

@@ -0,0 +1,62 @@
+#!/usr/bin/env python
+from tabulate import tabulate
+import algorithm
+import database
+import analysis
+import numpy as np
+
+
+def base_test(protocol, variables, setname='csh101', nb_tree_per_forest=50, max_depth=10):
+    # get train data
+    train = database.get(protocol, 'train', database.CLASSES, variables, setname)
+    # make and train model
+    model = algorithm.Model(nb_tree_per_forest, max_depth)
+    model.train(train)
+    # get test data
+    test = database.get(protocol, 'test', database.CLASSES, variables, setname)
+    test_labels = algorithm.make_labels(test).astype(int)
+    # make prediction on test
+    test_predictions = model.predict(test)
+    # get and return confusion matrix
+    cm = analysis.get_confusion_matrix(test_predictions, test_labels)
+    return cm
+
+
+def pretty_confusion_matrix(cm):
+    classes = np.array([database.CLASSES])
+    table = tabulate(np.vstack((np.hstack(([[""]], classes)),
+                                np.hstack((classes.T, cm)))))
+    return table
+
+
+def test_impact_nb_trees(tabnum):
+    nb_trees = [1, 5, 10]
+    print("\nImpact of number of trees per forest")
+    for n, p in enumerate(database.PROTOCOLS):
+        for m, nb_tree_per_forest in enumerate(nb_trees):
+            print("\nTable {table_number}: Confusion matrix with {nb_trees} tree(s) for Protocol `{protocol}`".format(
+                table_number=(n * len(nb_trees)) + m + tabnum,
+                protocol=p,
+                nb_trees=nb_tree_per_forest)
+            )
+            cm = base_test(p, database.VARIABLES, nb_tree_per_forest=nb_tree_per_forest)
+            print(pretty_confusion_matrix(cm))
+
+def test_impact_tree_depth(tabnum):
+    depths = [1, 5, 10]
+    print("\nImpact of trees maximum depth")
+    for n, p in enumerate(database.PROTOCOLS):
+        for m, max_depth in enumerate(depths):
+            print("\nTable {table_number}: Confusion matrix with trees maximum depth of {max_depth} for Protocol `{protocol}`".format(
+                table_number=(n * len(depths)) + m + tabnum,
+                protocol=p,
+                max_depth=max_depth)
+            )
+            cm = base_test(p, database.VARIABLES, max_depth=max_depth, nb_tree_per_forest=10)
+            print(pretty_confusion_matrix(cm))
+
+
+if __name__ == '__main__':
+    print("Main script for Human Activity Recognition with Random Forest classifier")
+    test_impact_nb_trees(1)
+    test_impact_tree_depth(7)