linreg-scratch 0.1.0__py3-none-any.whl

linreg_scratch/__init__.py

@@ -0,0 +1,4 @@
+from .model import LinearRegression
+
+__all__ = ["LinearRegression"]
+__version__ = "0.1.0"

linreg_scratch/model.py

@@ -0,0 +1,91 @@
+"""Simple linear regression (y = m*x + b) implemented from scratch with NumPy."""
+import numpy as np
+
+
+class LinearRegression:
+    """Simple (one feature) linear regression.
+
+    Parameters
+    ----------
+    lr : float
+        Learning rate for gradient descent.
+    epochs : int
+        Number of gradient descent iterations.
+    solver : {"gd", "closed"}
+        "gd" uses gradient descent; "closed" uses the exact least-squares formula.
+
+    Attributes
+    ----------
+    m : float
+        Slope.
+    b : float
+        Intercept (bias).
+    history : list of float
+        MSE at each epoch (gradient descent only).
+    """
+
+    def __init__(self, lr=0.01, epochs=1000, solver="gd"):
+        if solver not in ("gd", "closed"):
+            raise ValueError("solver must be 'gd' or 'closed'")
+        self.lr = lr
+        self.epochs = epochs
+        self.solver = solver
+        self.m = 0.0
+        self.b = 0.0
+        self.history = []
+        self._fitted = False
+
+    @staticmethod
+    def _to_1d(a, name):
+        a = np.asarray(a, dtype=float).ravel()
+        if a.size == 0:
+            raise ValueError(f"{name} must not be empty")
+        return a
+
+    def fit(self, x, y):
+        x = self._to_1d(x, "x")
+        y = self._to_1d(y, "y")
+        if x.shape != y.shape:
+            raise ValueError("x and y must have the same length")
+        n = len(x)
+
+        if self.solver == "closed":
+            x_mean, y_mean = x.mean(), y.mean()
+            var = np.sum((x - x_mean) ** 2)
+            if var == 0:
+                raise ValueError("x has zero variance; cannot fit a line")
+            self.m = float(np.sum((x - x_mean) * (y - y_mean)) / var)
+            self.b = float(y_mean - self.m * x_mean)
+        else:
+            self.m, self.b = 0.0, 0.0
+            self.history = []
+            for _ in range(self.epochs):
+                error = y - (self.m * x + self.b)
+                d_m = (-2.0 / n) * np.sum(x * error)
+                d_b = (-2.0 / n) * np.sum(error)
+                self.m -= self.lr * d_m
+                self.b -= self.lr * d_b
+                self.history.append(float(np.mean(error ** 2)))
+
+        self._fitted = True
+        return self
+
+    def predict(self, x):
+        if not self._fitted:
+            raise RuntimeError("Model is not fitted yet. Call fit() first.")
+        return self.m * np.asarray(x, dtype=float) + self.b
+
+    def mse(self, x, y):
+        y = self._to_1d(y, "y")
+        return float(np.mean((y - self.predict(self._to_1d(x, "x"))) ** 2))
+
+    def score(self, x, y):
+        """R^2 coefficient of determination."""
+        y = self._to_1d(y, "y")
+        pred = self.predict(self._to_1d(x, "x"))
+        ss_res = np.sum((y - pred) ** 2)
+        ss_tot = np.sum((y - y.mean()) ** 2)
+        return float(1 - ss_res / ss_tot) if ss_tot != 0 else 0.0
+
+    def __repr__(self):
+        return f"LinearRegression(m={self.m:.4f}, b={self.b:.4f}, solver='{self.solver}')"

linreg_scratch-0.1.0.dist-info/METADATA

@@ -0,0 +1,51 @@
+Metadata-Version: 2.4
+Name: linreg-scratch
+Version: 0.1.0
+Summary: Simple linear regression from scratch using only NumPy
+Author-email: Your Name <you@example.com>
+License: MIT
+Classifier: Programming Language :: Python :: 3
+Classifier: License :: OSI Approved :: MIT License
+Classifier: Operating System :: OS Independent
+Requires-Python: >=3.8
+Description-Content-Type: text/markdown
+License-File: LICENSE
+Requires-Dist: numpy>=1.20
+Dynamic: license-file
+
+# linreg-scratch
+
+Simple linear regression (`y = m*x + b`) implemented from scratch with NumPy.
+
+## Install
+
+```bash
+pip install linreg-scratch
+```
+
+## Usage
+
+```python
+import numpy as np
+from linreg_scratch import LinearRegression
+
+x = np.array([1, 2, 3, 4, 5])
+y = np.array([3, 5, 7, 9, 11])
+
+model = LinearRegression(lr=0.01, epochs=5000)   # gradient descent
+model.fit(x, y)
+
+print(model.m, model.b)       # slope, intercept
+print(model.predict([6, 7]))  # predictions
+print(model.score(x, y))      # R^2
+
+# exact solution instead of gradient descent
+LinearRegression(solver="closed").fit(x, y)
+```
+
+## Development
+
+```bash
+pip install -e . pytest
+pytest
+```

linreg_scratch-0.1.0.dist-info/RECORD

@@ -0,0 +1,7 @@
+linreg_scratch/__init__.py,sha256=TWJhZmv9UP2V0xcb66DlDmJb8ka6xH0X8QQz6bCNT14,90
+linreg_scratch/model.py,sha256=JT7gmKNVYwI0jOQfhwKJkIN8u2qP1k5bmZ3G-Gw0tLw,2939
+linreg_scratch-0.1.0.dist-info/licenses/LICENSE,sha256=v2spsd7N1pKFFh2G8wGP_45iwe5S0DYiJzG4im8Rupc,1066
+linreg_scratch-0.1.0.dist-info/METADATA,sha256=W7DvOlolPVAbwMnin5sCEHSGi_NhDiJ8Oy7H3TXUSt8,1121
+linreg_scratch-0.1.0.dist-info/WHEEL,sha256=mOdXR9C3wMVov7cSY8PY-XCzo3SmmWCHvGQr5csJoao,91
+linreg_scratch-0.1.0.dist-info/top_level.txt,sha256=xZZF9E0leQErfcqmtp0cyTH_CtLAAHlzQkSlDckWDBo,15
+linreg_scratch-0.1.0.dist-info/RECORD,,

linreg_scratch-0.1.0.dist-info/WHEEL

@@ -0,0 +1,5 @@
+Wheel-Version: 1.0
+Generator: setuptools (84.2.1)
+Root-Is-Purelib: true
+Tag: py3-none-any
+

linreg_scratch-0.1.0.dist-info/licenses/LICENSE

@@ -0,0 +1,21 @@
+MIT License
+
+Copyright (c) 2026 Your Name
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.

linreg_scratch-0.1.0.dist-info/top_level.txt

@@ -0,0 +1 @@
+linreg_scratch