webmining u4 2.2 confusuon matrix, accuracy(18%),precision, avg_precision

ss2013/1_Web Mining/Uebungen/4_Uebung/code/confusion_matrix.py

@@ -0,0 +1,85 @@
+import os
+
+actualDir = os.path.dirname(os.path.realpath(__file__))
+dataDir   = os.path.join(actualDir, '../data')
+trainDir  = os.path.join(dataDir, 'u4_train')
+testDir   = os.path.join(dataDir, 'u4_test')
+predfile  = os.path.join(actualDir, '../G22_predictions.txt')
+classes   = [['adventure',0],['belles_lettres',1],['editorial',2],['fiction',3],['government',4],['hobbies',5],['learned',6],['lore',7],['mystery',8],['news',9],['romance',10]]
+classes_n = 11
+
+preds      = []
+#each [0,0,0,0,0,0,0,0,0,0,0] = 1 column -> 1 column = adventure
+conf_matr = [[0,0,0,0,0,0,0,0,0,0,0],[0,0,0,0,0,0,0,0,0,0,0],[0,0,0,0,0,0,0,0,0,0,0],[0,0,0,0,0,0,0,0,0,0,0],[0,0,0,0,0,0,0,0,0,0,0],[0,0,0,0,0,0,0,0,0,0,0],[0,0,0,0,0,0,0,0,0,0,0],[0,0,0,0,0,0,0,0,0,0,0],[0,0,0,0,0,0,0,0,0,0,0],[0,0,0,0,0,0,0,0,0,0,0],[0,0,0,0,0,0,0,0,0,0,0]]            
+accuracy = 0
+document_count = 0
+precision = [0,0,0,0,0,0,0,0,0,0,0]
+precision_avg = 0
+
+def load_predictionfile():
+    global document_count
+    for line in open(predfile,'r'):
+        document_count += 1
+        str = line.split("\t")
+        preds.append([str[0],str[1].split("\n")[0]])
+    
+def confusion_matrix():     
+    for pred in preds:
+        for c in classes:
+            if os.path.isfile(trainDir +"/"+ c[0]+"/"+ pred[0]):
+                for p in classes:
+                    if pred[1] == p[0]:
+                        conf_matr[p[1]][c[1]] += 1            
+                
+def accuracy():
+    global accuracy
+    global document_count
+    i = 0
+    ok_recognized = 0
+    for conf in conf_matr:
+        ok_recognized += conf[i]
+        i += 1
+
+    #print ok_recognized
+    #print document_count        
+    accuracy = float(ok_recognized) / float(document_count)
+    
+def prec():
+    i = 0
+    global precision
+    for conf in conf_matr:
+        j = 0
+        ok_values = 0
+        not_ok_values = 0
+        for c in conf:
+            if j == i:
+                ok_values = c
+            else:
+                not_ok_values += c
+            j += 1
+        if not_ok_values +ok_values > 0:
+            precision[i] = float(ok_values) / float(ok_values + not_ok_values)
+        else:
+            precision[i] = 0
+        i += 1
+
+def avg_prec():
+    global precision_avg
+    global precision
+    for p in precision:
+        precision_avg += p
+    precision_avg = float(precision_avg) / 11
+        
+
+if __name__ == '__main__':
+    load_predictionfile()
+    #print preds
+    confusion_matrix()
+    #print conf_matr
+    accuracy()
+    #print "Accuracy: "+str(accuracy)
+    prec()
+    print precision
+    avg_prec()
+    print precision_avg
+