Fixes for Presentation

.gitignore

@@ -1 +1,2 @@
 __pycache__
+.venv

requirements

@@ -1,3 +1,4 @@
 matplotlib
 pandas
 tqdm
+scikit-learn

src/app.py

@@ -25,7 +25,7 @@ def auto_mpg() -> tuple[Dataset, MLAlgorithm, Any]:
     ds.numbers(["HP"])
     ds.handle_na()
     ds.normalize(excepts=["Cylinders","Year","Origin"])
-    return (ds, LinearRegression(ds, learning_rate=0.0001), sklearn.linear_model.LinearRegression())
+    return (ds, LinearRegression(ds, learning_rate=0.0001), sklearn.linear_model.SGDRegressor())
 
 def automobile() -> tuple[Dataset, MLAlgorithm, Any]:
     ds = Dataset(REGRESSION + "automobile.csv", "symboling", TargetType.Regression)
@@ -35,12 +35,12 @@ def automobile() -> tuple[Dataset, MLAlgorithm, Any]:
     ds.numbers(["normalized-losses", "bore", "stroke", "horsepower", "peak-rpm", "price"])
     ds.handle_na()
     ds.normalize(excepts=attributes_to_modify)
-    return (ds, LinearRegression(ds, learning_rate=0.004), sklearn.linear_model.LinearRegression())
+    return (ds, LinearRegression(ds, learning_rate=0.003), sklearn.linear_model.SGDRegressor())
 
 def power_plant() -> tuple[Dataset, MLAlgorithm, Any]:
     ds = Dataset(REGRESSION + "power-plant.csv", "energy-output", TargetType.Regression)
-    ds.normalize()
+    ds.normalize(excepts=None)
-    return (ds, LinearRegression(ds, learning_rate=0.1), sklearn.linear_model.LinearRegression())
+    return (ds, LinearRegression(ds, learning_rate=0.1), sklearn.linear_model.SGDRegressor())
 
 # ********************
 # Logistic Regression
@@ -101,7 +101,7 @@ def iris_no_target() -> tuple[Dataset, MLAlgorithm, Any]:
 if __name__ == "__main__":
     np.set_printoptions(linewidth=np.inf, formatter={'float': '{:>10.5f}'.format})
     rand = np.random.randint(0, 4294967295)
-    #rand = 1997847910  # LiR for power_plant
+    #rand = 2205910060  # LiR for power_plant
     #rand = 347617386   # LoR for electrical_grid
     #rand = 1793295160  # MLP for iris
     #rand = 2914000170  # MLP for frogs
@@ -110,7 +110,7 @@ if __name__ == "__main__":
     np.random.seed(rand)
     print(f"Using seed: {rand}")
 
-    ds, ml, sk = frogs()
+    ds, ml, sk = power_plant()
 
     epochs, _, _ = ml.learn(1000, verbose=True)
     ml.display_results()

src/learning/data.py

@@ -52,10 +52,8 @@ class Dataset:
 
         for col in self.data:
             if col not in excepts:
-                index = self.data.columns.get_loc(col)
+                datacol = self.data[col]
-                datacol = self.data.pop(col)
+                self.data[col] = (datacol - datacol.mean()) / datacol.std()
-                datacol = (datacol - datacol.mean()) / datacol.std()
-                self.data.insert(index, col, datacol)
         return self
 
     def factorize(self, columns:list[str]=[]) -> Self:

src/learning/functions.py

@@ -44,6 +44,13 @@ def cross_entropy_loss(h0:np.ndarray, y:np.ndarray) -> float:
 # Randoms
 # **********
 
+def pearson(h0:np.ndarray, y:np.ndarray) -> float:
+    diff1 = h0 - h0.mean()
+    diff2 = y - y.mean()
+    num = np.sum(diff1 * diff2)
+    den = np.sqrt(np.sum(diff1**2)) * np.sqrt(np.sum(diff2**2))
+    return num / den
+
 def r_squared(h0:np.ndarray, y:np.ndarray) -> float:
     y_mean = np.mean(y)
     ss_resid = np.sum((y - h0) ** 2)

src/learning/ml.py

@@ -5,7 +5,7 @@ from abc import ABC, abstractmethod
 from plot import Plot
 from tqdm import tqdm
 from learning.data import ConfusionMatrix, Dataset, Data, TargetType
-from learning.functions import r_squared
+from learning.functions import pearson, r_squared
 
 class MLAlgorithm(ABC):
     """ Classe generica per gli algoritmi di Machine Learning """
@@ -83,6 +83,7 @@ class MLAlgorithm(ABC):
         print(f"Loss valid : {self.validation_loss():0.5f}")
         print(f"Loss test  : {self.test_loss():0.5f}")
         if self._target_type == TargetType.Regression:
+            print(f"Pearson    : {self.test_pearson():0.5f}")
             print(f"R^2        : {self.test_r_squared():0.5f}")
         elif self._target_type != TargetType.NoTarget:
             conf = self.test_confusion_matrix()
@@ -106,6 +107,11 @@ class MLAlgorithm(ABC):
             y = np.argmax(y, axis=1)
         return ConfusionMatrix(y, h0)
 
+    def test_pearson(self) -> float:
+        if self._target_type != TargetType.Regression:
+            return 0
+        return pearson(self._h0(self._testset.x), self._testset.y)
+
     def test_r_squared(self) -> float:
         if self._target_type != TargetType.Regression:
             return 0

src/learning/supervised.py

@@ -19,11 +19,13 @@ class GradientDescent(MLAlgorithm):
 
     def _learning_step(self) -> float:
         x, y, m, _ = self._learnset.as_tuple()
+        h0 = self._h0(x)
 
         regularization = (self.lambd / m) * self.theta
         regularization[0] = 0
-        derivative = self.alpha * np.mean((self._h0(x) - y) * with_bias(x).T, axis=1)
+
-        self.theta -= derivative + regularization
+        derivative =  np.mean((h0 - y) * with_bias(x).T, axis=1)
+        self.theta -= self.alpha * derivative + regularization
         return self._loss(x, y)
 
     def _predict_loss(self, dataset:Data) -> float: