print("Nominal dataset")
# Create a Naive Bayes Classifier supporting both numerical
# and categorical features
# Load categorical data using code from the previous exercise
X_flu_train, y_flu_train, X_flu_test = load_flu()
# Implement and test Naive Bayes Classifier supporting categorical data
nb = NaiveBayesNumNom()
nb.fit(X_flu_train, y_flu_train)
# Use the previously implemented testing function
# to evaluate the performance of the model
test_model(nb, X_flu_train, y_flu_train)
test_model(nb, X_flu_test)
# ground truth:
# [0 1 1 1 0 1 0 1]
# predicted:
# [1 1 1 1 0 1 0 1]
# accuracy:
# 0.875
# predicted:
# [1]
############################################################################################################################################################
print("Numerical dataset")
y_iris = iris.target # Divide the dataset into training and test part 80:20 respectively frac_train = 0.8 n_samples = X_iris.shape[0] n_train = int(float(n_samples * frac_train)) X_iris_train = X_iris[:n_train, :] y_iris_train = y_iris[:n_train] X_iris_test = X_iris[n_train:, :] y_iris_test = y_iris[n_train:] # Implement and test Naive Bayes Classifier supporting categorical data mgnb = NaiveBayesGaussian() mgnb.fit(X_iris_train, y_iris_train) test_model(mgnb, X_iris_train, y_iris_train, False) test_model(mgnb, X_iris_test, y_iris_test, False) # accuracy: # 0.966666666667 # accuracy: # 0.933333333333 # Compare the performance of your implementation # to GaussianNB from sklearn to verify correctness # from sklearn.naive_bayes import GaussianNB # # gnb = GaussianNB() # gnb.fit(X_iris_train, y_iris_train) # # test_model(gnb, X_iris_train, y_iris_train, False)
''' Author: Kalina Jasinska ''' from utils.load import load_flu from classifiers_students.naive_bayes import NaiveBayesNominal from utils.evaluate import test_model # Load categorical data using code from the previous exercise X_flu_train, y_flu_train, X_flu_test = load_flu() # Implement and test Naive Bayes Classifier supporting categorical data nb = NaiveBayesNominal() nb.fit(X_flu_train, y_flu_train) # Use the previously implemented testing function # to evaluate the performance of the model test_model(nb, X_flu_train, y_flu_train) test_model(nb, X_flu_test) # ground truth: # [0 1 1 1 0 1 0 1] # predicted: # [1 1 1 1 0 1 0 1] # accuracy: # 0.875 # predicted: # [1]