Revisions

models/features.py

@@ -0,0 +1,51 @@
+from sklearn.feature_extraction import DictVectorizer
+from sklearn.linear_model import LogisticRegression
+from sklearn.pipeline import Pipeline
+
+
+class FeatureExtractor:
+    def fit(self, X, y=None):
+        return self
+
+    def transform(self, X):
+        return [self.extract_features(pixels) for pixels in X]
+
+    def extract_features(self, pixels):
+        """Extract hand-designed features from a 784-pixel image.
+
+        Add at least two features of your own. Each feature should be a
+        number computed from the pixel array.
+
+        Arguments:
+            pixels: numpy array of 784 float values in [0, 1]
+
+        Returns:
+            dict: feature name -> numerical value
+        """
+        img = pixels.reshape(28, 28)
+        return {
+            "mean_brightness": float(pixels.mean()),
+            "top_half_brightness": float(img[:14, :].mean()),
+            "bottom_half_brightness": float(img[14:, :].mean()),
+            # ---- Add your features here ----
+            # "left_half_brightness": float(img[:, :14].mean()),
+            # "right_half_brightness": float(img[:, 14:].mean()),
+            # "num_bright_pixels": float((pixels > 0.5).sum()),
+        }
+
+
+class FeatureClassifier:
+    def fit(self, X, y):
+        self._pipeline = Pipeline([
+            ("features", FeatureExtractor()),
+            ("vectorizer", DictVectorizer()),
+            ("classifier", LogisticRegression(max_iter=1000)),
+        ])
+        self._pipeline.fit(X, y)
+        return self
+
+    def predict(self, X):
+        return self._pipeline.predict(X)
+
+    def predict_proba(self, X):
+        return self._pipeline.predict_proba(X)