Added Final version & report

ImageClassification/Finalversion.py

@@ -0,0 +1,256 @@
+
+# coding: utf-8
+
+# In[75]:
+
+
+import pandas as pd
+import numpy as np
+import operator
+
+from sklearn.preprocessing import StandardScaler
+
+from sklearn.pipeline import make_pipeline
+from xgboost import XGBClassifier
+
+from sklearn.model_selection import train_test_split
+from imblearn.over_sampling import SMOTE
+from sklearn.preprocessing import StandardScaler, RobustScaler, MinMaxScaler
+
+
+from collections import Counter
+from sklearn.metrics import confusion_matrix,classification_report
+from sklearn.model_selection import cross_val_score
+import warnings
+import time
+
+import matplotlib.pyplot as plt
+from sklearn import svm
+
+from sklearn.feature_selection import VarianceThreshold
+
+from sklearn.tree import DecisionTreeClassifier
+from sklearn.neighbors import KNeighborsClassifier
+from sklearn.svm import SVC
+from sklearn.ensemble import RandomForestClassifier, ExtraTreesClassifier, VotingClassifier, AdaBoostClassifier
+from sklearn.ensemble import BaggingClassifier
+
+from mlxtend.classifier import StackingClassifier
+
+from sklearn.multiclass import OneVsRestClassifier
+from imblearn.combine import SMOTEENN
+from imblearn.over_sampling import SMOTE
+from imblearn.combine import SMOTETomek
+
+
+# In[2]:
+
+
+pd.set_option('max_colwidth', 1000)
+
+trainrecord = pd.read_csv('/Users/whiplash/SJSU/Semester 2/CMPE 255/Assignments/Program 2/data/train.txt', 
+                          header=None, delimiter=' ')
+
+trainlabels = pd.read_csv('/Users/whiplash/SJSU/Semester 2/CMPE 255/Assignments/Program 2/data/train.labels', 
+                          header=None, names =['labels'])
+
+testrecord = pd.read_csv('/Users/whiplash/SJSU/Semester 2/CMPE 255/Assignments/Program 2/data/test.txt', 
+                          header=None, delimiter=' ')
+
+
+# In[77]:
+
+
+target = []
+for x, value in np.ndenumerate(trainlabels):
+    target.append(value)
+
+X = trainrecord
+y = target
+Xtestfinal = testrecord
+
+
+# In[78]:
+
+
+print(type(X))
+
+
+# In[79]:
+
+
+dataset = []
+labels = []
+for row, x in enumerate(y):
+    if y[row] == 1:
+        dataset.append(X.iloc[row])
+        labels.append(y[row])
+    if y[row] == 2:
+        dataset.append(X.iloc[row])
+        labels.append(y[row])
+    if y[row] == 3:
+        dataset.append(X.iloc[row])
+        labels.append(y[row])
+    if y[row] == 4:
+        dataset.append(X.iloc[row])
+        labels.append(y[row])
+    if y[row] == 5:
+        dataset.append(X.iloc[row])
+        labels.append(y[row])
+    if y[row] == 6:
+        dataset.append(X.iloc[row])
+        labels.append(y[row])
+    if y[row] == 7:
+        dataset.append(X.iloc[row])
+        labels.append(y[row])
+    if y[row] == 8:
+        dataset.append(X.iloc[row])
+        labels.append(y[row])
+
+
+# In[80]:
+
+
+Counter(labels)
+
+
+# In[85]:
+
+
+scaler = StandardScaler()
+X_scaled = scaler.fit_transform(X)
+
+
+# In[86]:
+
+
+def feature_selection(train_instances):
+    print('Crossvalidation started... ')
+    selector = VarianceThreshold()
+    selector.fit(train_instances)
+    print('Number of features used... ' + str(Counter(selector.get_support())[True]))
+    print('Number of features ignored... ' +str(Counter(selector.get_support())[False]))
+    return selector
+
+
+# In[87]:
+
+
+#Learn the features to filter from train set
+print("Selecting features... ")
+fs = feature_selection(X_scaled)
+
+#Transform train and test subsets
+train_instances = fs.transform(X_scaled)
+
+
+# In[89]:
+
+
+sm = SMOTE(ratio = {5:2000, 11:2000, 10:2000, 6:2000, 7:2000, 4:2000, 8:2000}, random_state=42, k_neighbors = 2)
+X_res, y_res = sm.fit_sample(train_instances, y)
+
+
+# In[90]:
+
+
+Counter(y_res)
+
+
+# In[91]:
+
+
+def evaluate(classifier, training_instances, training_labels):
+    metrics = cross_validate(classifier, training_instances, training_labels, cv=10, 
+                             n_jobs=-1, scoring=['accuracy'])
+    print("Accuracy: %0.4f (+/- %0.4f)" % (metrics['test_accuracy'].mean(), metrics['test_accuracy'].std() * 2))
+    print("Mean fit time: %0.4f ms" % (metrics['fit_time'].mean()))
+
+
+# In[101]:
+
+
+voting = VotingClassifier([ 
+    ('knn', OneVsRestClassifier(KNeighborsClassifier(n_neighbors = 4))),
+    ('et', OneVsRestClassifier(ExtraTreesClassifier(n_estimators = 1000))),
+    ('xgboost', XGBClassifier(learning_rate =0.1, n_estimators=1000, num_class = 12,
+                             min_child_weight=1, gamma=0,subsample=0.8,colsample_bytree=0.8,
+                             objective= 'multi:softmax', nthread=4,scale_pos_weight=1,
+                             seed=27, early_stopping_rounds=70, verbose=False)),
+    ('randomforest', RandomForestClassifier(n_estimators=1000)),
+    ('ada', AdaBoostClassifier(n_estimators=1000))], voting='soft', weights =[1,2,1,2,1])
+
+
+# In[102]:
+
+
+X_test_scaled = scaler.transform(Xtestfinal)
+test_instances = fs.transform(X_test_scaled)
+
+
+# In[103]:
+
+
+voting.fit(X_res, y_res)
+
+
+# In[104]:
+
+
+predictions = voting.predict(test_instances)
+
+
+# In[105]:
+
+
+Counter(predictions)
+
+
+# In[106]:
+
+
+predictionnew = predictions
+
+
+# In[107]:
+
+
+knn = KNeighborsClassifier(n_neighbors = 4)
+rf = RandomForestClassifier(n_estimators = 1000)
+
+
+# In[98]:
+
+
+knn.fit(X_res, y_res)
+rf.fit(X_res, y_res)
+
+
+# In[99]:
+
+
+knnp = knn.predict(test_instances)
+rfp = rf.predict(test_instances)
+
+
+# In[108]:
+
+
+newlist = []
+get_indexes = lambda knnp, xs: [i for (y, i) in zip(xs, range(len(xs))) if knnp == y]
+newlist.append(get_indexes(10, knnp))
+for item in newlist:
+    predictionnew[item] = 10
+newlist = []
+get_indexes = lambda rfp, xs: [i for (y, i) in zip(xs, range(len(xs))) if rfp == y]
+newlist.append(get_indexes(7, rfp))
+for item in newlist:
+    predictionnew[item] = 7
+Counter(predictionnew)
+
+
+# In[109]:
+
+
+np.savetxt('predictions.dat', predictionnew, delimiter=" ", fmt="%s")
+