Add simple linear regression implementation using OOP concepts

linear_regression.py

@@ -0,0 +1,106 @@
+# Python Object Oriented Programming by Joe Marini course example
+# Simple linear regression using OOP concepts
+
+from dataclasses import dataclass
+
+
+@dataclass
+class DataPoint:
+    """Represents a single (x, y) data observation."""
+    x: float
+    y: float
+
+    def __str__(self):
+        return f"({self.x}, {self.y})"
+
+
+class SimpleLinearRegression:
+    """Performs simple linear regression (y = mx + b) using least squares."""
+
+    def __init__(self, data_points):
+        if len(data_points) < 2:
+            raise ValueError("At least 2 data points are required")
+
+        self._x_data = [p.x for p in data_points]
+        self._y_data = [p.y for p in data_points]
+        self._slope = None
+        self._intercept = None
+        self._fit()
+
+    def _fit(self):
+        """Compute slope and intercept using the least-squares method."""
+        n = len(self._x_data)
+        sum_x = sum(self._x_data)
+        sum_y = sum(self._y_data)
+        sum_xy = sum(x * y for x, y in zip(self._x_data, self._y_data))
+        sum_x2 = sum(x ** 2 for x in self._x_data)
+
+        denominator = n * sum_x2 - sum_x ** 2
+        if denominator == 0:
+            raise ValueError("All x values are identical; cannot compute regression")
+
+        self._slope = (n * sum_xy - sum_x * sum_y) / denominator
+        self._intercept = (sum_y - self._slope * sum_x) / n
+
+    def predict(self, x):
+        """Predict y for a given x value."""
+        return self._slope * x + self._intercept
+
+    def get_r_squared(self):
+        """Compute the R-squared (coefficient of determination)."""
+        y_mean = sum(self._y_data) / len(self._y_data)
+        ss_tot = sum((y - y_mean) ** 2 for y in self._y_data)
+        if ss_tot == 0:
+            return 1.0
+        ss_res = sum(
+            (y - self.predict(x)) ** 2
+            for x, y in zip(self._x_data, self._y_data)
+        )
+        return 1 - (ss_res / ss_tot)
+
+    def __str__(self):
+        return (f"LinearRegression: y = {self._slope:.4f}x + "
+                f"{self._intercept:.4f} (R²={self.get_r_squared():.4f})")
+
+    def __repr__(self):
+        return (f"SimpleLinearRegression(n={len(self._x_data)}, "
+                f"slope={self._slope:.4f}, intercept={self._intercept:.4f})")
+
+    def __call__(self, x):
+        """Make the model callable for predictions."""
+        return self.predict(x)
+
+
+# ~~~~~~~~~ TEST CODE ~~~~~~~~~
+# Sample data: years of experience vs. salary (in thousands)
+data = [
+    DataPoint(1.0, 45.0),
+    DataPoint(2.0, 50.0),
+    DataPoint(3.0, 55.0),
+    DataPoint(4.0, 60.0),
+    DataPoint(5.0, 68.0),
+    DataPoint(6.0, 72.0),
+    DataPoint(7.0, 80.0),
+    DataPoint(8.0, 85.0),
+    DataPoint(9.0, 90.0),
+    DataPoint(10.0, 95.0),
+]
+
+# Create the regression model
+model = SimpleLinearRegression(data)
+
+# Display the model equation
+print(model)
+
+# Make predictions using the predict method
+print(f"Predicted salary for 5 years exp:  {model.predict(5.0):.2f}k")
+print(f"Predicted salary for 15 years exp: {model.predict(15.0):.2f}k")
+
+# Use the model as a callable (demonstrates __call__)
+print(f"Predicted salary for 12 years exp: {model(12.0):.2f}k")
+
+# Display model details using repr
+print(repr(model))
+
+# Show R-squared value
+print(f"R-squared: {model.get_r_squared():.4f}")