Linear regression higher F1.0 lower accuracy

predict.py

@@ -8,6 +8,8 @@ from tokenizer import tokenize
 #Load model
 model = pickle.load(open("model.pkl","rb"))
 weights, word_to_index_mapping, word_count = model
+sum = 0
+counter = 0
 
 for line in sys.stdin:
     document = line.rstrip()
@@ -17,9 +19,12 @@ for line in sys.stdin:
 
     y_predicted = weights[0]
     for word in terms:
-        y_predicted += weights[word_to_index_mapping.get(word,0)] * (word_count.get(word,0) / len(word_count))
-
-    if y_predicted <= 0.5:
+        y_predicted += weights[word_to_index_mapping.get(word,0)] * log(word_count.get(word,0) / len(word_count) + 1)
+    sum += y_predicted
+    counter += 1
+    if y_predicted <= 0:
         print(0)
     else:
-        print(1)
+        print(1)
+
+#print(sum / counter)

train.py

@@ -76,7 +76,7 @@ def train():
         #With get u can avoid missing words and replace them with value u want
         #Weights replace value doesnt matter if word is missing cause word_count will give 0
         for word in actual_x:
-            y_predicted += weights[word_to_index_mapping.get(word,0)] * (word_count.get(word,0) / len(word_count))
+            y_predicted += weights[word_to_index_mapping.get(word,0)] * log(word_count.get(word,0) / len(word_count) + 1)
 
         #Cost count. Check how good was our prediction
         Loss = (y_predicted - actual_y) ** 2.0
@@ -94,9 +94,9 @@ def train():
         weights[0] = weights[0] - delta
         for word in actual_x:
             if word in word_to_index_mapping:
-                weights[word_to_index_mapping[word]] -= ((word_count[word] / len(word_count)) * delta)
+                weights[word_to_index_mapping[word]] -= (log(word_count[word] / len(word_count) + 1) * delta)
 
-        if Loss_sum_counter > 10000000:
+        if Loss_sum_counter > 7000000:
             break