From 361ef5950099e954ea625c8df103f3dc127b005b Mon Sep 17 00:00:00 2001
From: Xi Xu <i@xi-xu.me>
Date: Wed, 6 Nov 2024 19:13:43 +0800
Subject: [PATCH] Update LinearClassifier_1.py

---
 04_linear_classifier/LinearClassifier_1.py | 84 ++++++++++++----------
 1 file changed, 47 insertions(+), 37 deletions(-)

diff --git a/04_linear_classifier/LinearClassifier_1.py b/04_linear_classifier/LinearClassifier_1.py
index beabc50..3575d18 100644
--- a/04_linear_classifier/LinearClassifier_1.py
+++ b/04_linear_classifier/LinearClassifier_1.py
@@ -1,44 +1,49 @@
 import sys
-sys.path.insert(1, '../')
-from utils.utils import *
+
+sys.path.insert(1, "../")
 from utils.dataset4learners import *
+from utils.utils import *
 
 
 class LinearClassifier:
     def learn(self, learning_rate, epochs):
         raise NotImplementedError
-        
+
     def predict(self, x):
         raise NotImplementedError
-    
-    
+
+
 class PerceptionLinearLearner(LinearClassifier):
     """
     Perception linear classifier: hard threshold
     """
+
     def __init__(self, dataset, learning_rate=0.01, epochs=100):
         self.idx_i = dataset.inputs
         self.idx_t = dataset.target
         self.examples = dataset.examples
         self.num_examples = len(self.examples)
         # initialize random weights
-        self.w = random_weights(min_value=-0.5, max_value=0.5, num_weights=len(self.idx_i) + 1)
+        self.w = random_weights(
+            min_value=-0.5, max_value=0.5, num_weights=len(self.idx_i) + 1
+        )
         # learning loop
         self.learn(learning_rate, epochs)
-        
+
     def learn(self, learning_rate, epochs):
-        """ learning loop """
+        """learning loop"""
+
         def loss(example, w, idx_i, idx_t):
-            """ error: difference between estimation and true value """
+            """error: difference between estimation and true value"""
             raise NotImplementedError
-        
+
         def update(w, learning_rate, err, X_col, num_examples):
-            """ update weights """
+            """update weights"""
             for i in range(len(w)):
                 w[i] = w[i] - learning_rate * (np.dot(err, X_col[i]) / num_examples)
-                
+
         def homogeneous(num_examples):
-            """ build homogeneous coordinates """
+            """build homogeneous coordinates"""
             raise NotImplementedError
 
         X_col = homogeneous(self.num_examples)
@@ -52,7 +57,7 @@ class PerceptionLinearLearner(LinearClassifier):
             update(self.w, learning_rate, err, X_col, self.num_examples)
 
     def predict(self, x):
-        """ make prediction """
+        """make prediction"""
         return int(np.dot(self.w, [1] + x))
 
 
@@ -63,24 +68,27 @@ class LogisticLinearLeaner(LinearClassifier):
         self.examples = dataset.examples
         self.num_examples = len(self.examples)
         # initialize random weights
-        self.w = random_weights(min_value=-0.5, max_value=0.5, num_weights=len(self.idx_i) + 1)
+        self.w = random_weights(
+            min_value=-0.5, max_value=0.5, num_weights=len(self.idx_i) + 1
+        )
         # learning loop
         self.learn(learning_rate, epochs)
-        
+
     def learn(self, learning_rate, epochs):
-        """ learning loop """
+        """learning loop"""
+
         def loss(example, w, idx_i, idx_t, h):
-            """ error: difference between estimation and true value """
+            """error: difference between estimation and true value"""
             raise NotImplementedError
-        
+
         def update(w, learning_rate, err, h, X_col, num_examples):
-            """ update weights """
+            """update weights"""
             for i in range(len(w)):
                 buffer = [x * y for x, y in zip(err, h)]
                 w[i] = w[i] - learning_rate * (np.dot(buffer, X_col[i]) / num_examples)
 
         def homogeneous(num_examples):
-            """ build homogeneous coordinates """
+            """build homogeneous coordinates"""
             raise NotImplementedError
 
         X_col = homogeneous(self.num_examples)
@@ -95,35 +103,37 @@ class LogisticLinearLeaner(LinearClassifier):
             update(self.w, learning_rate, err, h, X_col, self.num_examples)
 
     def predict(self, x):
-        """ make prediction """
+        """make prediction"""
         return int(np.dot(self.w, [1] + x))
 
 
 if __name__ == "__main__":
     iris = DataSet(name="iris")
     iris.classes_to_numbers()
-    
-    tests = [([5, 3, 1, 0.1], 0),
-             ([5, 3.5, 1, 0], 0),
-             ([6, 3, 4, 1.1], 1),
-             ([6, 2, 3.5, 1], 1),
-             ([7.5, 4, 6, 2], 2),
-             ([7, 3, 6, 2.5], 2)]
-    
-    print(f'===================\nperceptron:')
+
+    tests = [
+        ([5, 3, 1, 0.1], 0),
+        ([5, 3.5, 1, 0], 0),
+        ([6, 3, 4, 1.1], 1),
+        ([6, 2, 3.5, 1], 1),
+        ([7.5, 4, 6, 2], 2),
+        ([7, 3, 6, 2.5], 2),
+    ]
+
+    print(f"===================\nperceptron:")
     perceptron = PerceptionLinearLearner(iris)
     g = grade_learner(perceptron, tests)
-    print(f'  learner grade: {g}')
+    print(f"  learner grade: {g}")
     # assert g > 1. / 2
     e = err_ratio(perceptron, iris)
-    print(f'  error ration: {e}')
+    print(f"  error ration: {e}")
     # assert e < 0.4
-    
-    print(f'===================\nlogistic:')
+
+    print(f"===================\nlogistic:")
     logisticer = LogisticLinearLeaner(iris)
     g = grade_learner(logisticer, tests)
-    print(f'  learner grade: {g}')
+    print(f"  learner grade: {g}")
     # assert g > 1. / 2
     e = err_ratio(logisticer, iris)
-    print(f'  error ration: {e}')
+    print(f"  error ration: {e}")
     # assert e < 0.4