achieves ~10 MSE

driver.py

@@ -1,20 +1,26 @@
 #!/usr/bin/env python
+import numpy as np
+from sklearn import linear_model
 from model_factory_method import model_create
 from extract_features import prepare_features
 from IPython import embed
 from os import environ
+from sklearn.metrics import accuracy_score, precision_score
+
+seed = 7
+np.random.seed(seed)
 
 tests = [
 ["model_time",       "./dataset/NN_test_Y_ADO.mat", "sWave_ADO"], 
-#["model_time",       "./dataset/NN_test_Y_RPV.mat", "sWave_RPV"],
-#["model_time",       "./dataset/NN_test_Y_USC.mat", "sWave_USC"],
-#["model_time",       "./dataset/NN_test_Y_RSS.mat", "sWave_RSS"],
-#["model_depth",      "./dataset/NN_test_Y_depth.mat", "data_depth"],
-#["model_location",   "./dataset/NN_test_Y_eqLoc.mat", "data_eqLoc"],
-#["model_magnitude",  "./dataset/NN_test_Y_magnitude.mat", "data_mag"]
+["model_time",       "./dataset/NN_test_Y_RPV.mat", "sWave_RPV"],
+["model_time",       "./dataset/NN_test_Y_USC.mat", "sWave_USC"],
+["model_time",       "./dataset/NN_test_Y_RSS.mat", "sWave_RSS"],
+["model_depth",      "./dataset/NN_test_Y_depth.mat", "data_depth"],
+["model_magnitude",  "./dataset/NN_test_Y_magnitude.mat", "data_mag"],
+#["model_location",   "./dataset/NN_test_Y_eqLoc.mat", "data_eqLoc"]
 ]
 
-BATCH_SIZE = 1000
+BATCH_SIZE = 50
 
 EPOCHS     = 2000
 if 'EPOCHS' in environ:
@@ -26,19 +32,18 @@
     model_factory = model_create(setup[0],53)
     model = model_factory.model
 
-    #embed()
-
-    model.compile(loss='mean_squared_error', optimizer='adam', metrics=['accuracy'])
+    model.compile(loss='mean_squared_error', optimizer='adam', metrics=['mse','accuracy'])
 
     model.fit(train_x, train_y,
               batch_size=BATCH_SIZE,
               shuffle=True,
               nb_epoch=EPOCHS,
               verbose=1,
-              validation_data=(test_x, test_y))
+              validation_split=0.2)
+              #validation_data=(test_x, test_y))
 
     score = model.evaluate(test_x, test_y, verbose=0)
     print('Test loss:', score[0])
-    print('Test accuracy:', score[1])
+    print('Test mse:', score[1])
 
 embed()

extract_features.py

@@ -19,6 +19,6 @@ def prepare_features(labels_file, filekey):
     dataset_x = np.hstack((dataset_x, extra2))
     dataset_y = sio.loadmat(labels_file)[filekey]
 
-    train_x, test_x, train_y, test_y = train_test_split(dataset_x, dataset_y, test_size=0.4)
+    train_x, test_x, train_y, test_y = train_test_split(dataset_x, dataset_y, test_size=0.2)
 
     return train_x, train_y, test_x, test_y

model_depth.py

@@ -5,9 +5,10 @@
 
 class ModelDepth(ModelFactory):
     def create(self):
-        self.model = Sequential()
         model = Sequential()
-        model.add(Dense(46, input_shape=(self.dim,), activation='relu'))
-        model.add(Dense(512, activation='sigmoid'))
-        model.add(Dense(1, activation='softmax'))
+        model.add(Dense(53, init='normal', input_shape=(self.dim,), activation='relu'))
+        model.add(Dense(8, activation='relu'))
+        model.add(Dense(5, activation='relu'))
+        model.add(Dense(1, init='normal')) #, activation='elu'))
+        self.model = model
         return model

model_location.py

@@ -5,9 +5,11 @@
 
 class ModelLocation(ModelFactory):
     def create(self):
-        self.model = Sequential()
         model = Sequential()
-        model.add(Dense(46, input_shape=(self.dim,), activation='relu'))
-        model.add(Dense(512, activation='sigmoid'))
-        model.add(Dense(1, activation='softmax'))
+        model.add(Dense(53, init='normal', input_shape=(self.dim,), activation='relu'))
+        model.add(Dense(8, activation='relu'))
+        model.add(Dense(5, activation='relu'))
+        model.add(Dense(1, init='normal')) #, activation='elu'))
+        self.model = model
         return model
+

model_magnitude.py

@@ -5,9 +5,10 @@
 
 class ModelMagnitude(ModelFactory):
     def create(self):
-        self.model = Sequential()
         model = Sequential()
-        model.add(Dense(46, input_shape=(self.input_dim,), activation='relu'))
-        model.add(Dense(512, activation='sigmoid'))
-        model.add(Dense(1, activation='softmax'))
+        model.add(Dense(53, init='normal', input_shape=(self.dim,), activation='relu'))
+        model.add(Dense(8, activation='relu'))
+        model.add(Dense(5, activation='relu'))
+        model.add(Dense(1, init='normal')) #, activation='elu'))
+        self.model = model
         return model

model_time.py

@@ -5,13 +5,10 @@
 
 class ModelTime(ModelFactory):
     def create(self):
-        self.model = Sequential()
         model = Sequential()
-        model.add(Dense(53, input_shape=(self.dim,), activation='relu'))
-#        model.add(Dropout(0.25))
-        model.add(Dense(25))
-        #model.add(Dropout(0.25))
-        #model.add(Dense(20))
-        #model.add(Dense(10))
-        model.add(Dense(1)) #, activation='sigmoid'))
+        model.add(Dense(53, init='normal', input_shape=(self.dim,), activation='relu'))
+        model.add(Dense(8, activation='relu'))
+        model.add(Dense(5, activation='relu'))
+        model.add(Dense(1, init='normal')) #, activation='elu'))
+        self.model = model
         return model