Naive bayes Trees

from sklearn.naive_bayes import CategoricalNB
from sklearn.preprocessing import OrdinalEncoder
import math

if __name__ == '__main__':
    dataset = [
      ['high', 'light', 'hard', 'dab'],
      ['high', 'light', 'hard', 'bor'],
      ['high', 'light', 'hard', 'dab'],
      ['high', 'light', 'soft', 'bor'],
      ['high', 'dark', 'hard', 'dab'],
      ['high', 'dark', 'soft', 'bor'],
      ['low', 'dark', 'soft', 'dab'],
      ['high', 'light', 'soft', 'bor'],
      ['low', 'dark', 'hard', 'bor'],
      ['low', 'dark', 'hard', 'dab']
    ]
    encoder = OrdinalEncoder()
    encoder.fit([dataset[i][:-1] for i in range(0, len(dataset))])
    
    upperThreshold = math.ceil(0.67 * len(dataset))
    train_set = dataset[0:upperThreshold]
    test_set = dataset[upperThreshold:]
    
    X = encoder.transform([train_set[i][:-1] for i in range(0, len(train_set))])
    Y = [train_set[i][-1] for i in range(0, len(train_set))]
    
    clf = CategoricalNB(alpha=2.0)
    clf.fit(X, Y)
    
    test_set_x = encoder.transform([test_set[i][:-1] for i in range(0, len(test_set))])
    
    accuracy = 0
    for i in range(0,len(test_set)):
        predict = clf.predict([test_set_x[i]])
        if predict[0] == test_set[i][-1]:
            accuracy += 1
    print(accuracy / len(test_set))

    
    entry = encoder.transform([['low', 'light', 'hard']])
    print(clf.predict(entry))
    print(clf.predict_proba(entry))