Naive Bayes attempt to classify boolean data

naiveBayes.py

"""
Feature is 512 1-s or 0-s, label is 1 or 0 
Data represents some bio research on components added to a mixture,
all are independant one from another.
Outcome is passing some threshold for a further research.
"""

import numpy as np
from sklearn.naive_bayes import GaussianNB

def getData():
    features = np.reshape(
        np.fromfile("bytes/data600.dat", dtype=np.uint8),
        [-1,512]
    )
    labels = np.fromfile("bytes/labels600.dat", dtype=np.uint8)
    split = 599900
    return features[:split], labels[:split], features[split:], labels[split:]
    
# Initialize
Xtrain, Ytrain, Xtest, Ytest = getData()
naiveBayes = GaussianNB()

# Train
naiveBayes.fit(Xtrain, Ytrain)

# Make prediction
prediction = naiveBayes.predict(Xtest)

# Evaluate
mismatched = np.logical_xor(prediction, Ytest)
accuracy = (len(Ytest) - np.count_nonzero(mismatched)) / len(Ytest)
print("Accuracy:", accuracy)
print("Expectation:", Ytest[:20],"Reality:", prediction[:20])

"""
Accuracy: 0.47
Expectation: [1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0]
Reality: [1 0 0 1 1 1 1 0 0 0 0 1 0 1 1 1 1 1 1 1]
"""