A model can memorize the examples it receives, yet fail on tomorrow’s data. train_test_split randomly divides arrays into a training dataset for fitting and a test dataset for evaluating unseen cases. section 8.5 section 8.6
A practical split follows this pattern:
- Reserve part of the data for testing; a common ratio uses two-thirds for training and one-third for testing. section 8.6
- Pass the training portion to the model’s fit method, keeping the test portion out of the training process. section 8.6
- Use the fitted model to predict the test labels, then compare predictions with actual labels to calculate accuracy. section 8.6
- Treat the result as an estimate of generalization, not automatic proof; cross-validation is preferred when fitting is not computationally expensive. section 8.11
Predicting toddler meltdowns
For a binary meltdown example, train_test_split(x, y, test_size=1/3) randomly creates X_train, X_test, Y_train, and Y_test. LogisticRegression fits the training arrays, predicts classes on X_test, and compares those predictions with Y_test. The same pattern works with linear regression, logistic regression, or neural networks. section 8.6
import numpy as np
from sklearn.linear_model import LogisticRegression
from sklearn.model_selection import train_test_split
x = np.array([0.1, 0.5, 0.7, 1.2, 1.2, 1.5, 1.6, 1.9, 2.5, 2.5,
2.7, 3.2, 3.3, 3.7, 3.8, 4.1, 4.5, 4.5, 4.5, 5.0,
5.1, 5.6, 5.7, 6.2, 6.3]) \
.reshape(-1, 1) #1
y = np.array([0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 1.0,
0.0, 0.0, 1.0, 1.0, 0.0, 1.0, 1.0, 1.0, 1.0, 1.0,
1.0, 1.0, 1.0, 1.0, 1.0]) #2
X_train, X_test, Y_train, Y_test = \
train_test_split(x, y, test_size=1/3) #3
model = LogisticRegression(penalty=None) #4
model.fit(X_train, Y_train) #4
predictions = model.predict(X_test) #5
print(predictions) #6
accuracy = np.sum(predictions == Y_test) / len(Y_test) #7
print(f"Accuracy: {accuracy:.4f}") #7
Inspect the complete NumPy and scikit-learn listing: it creates the arrays, performs the split, fits logistic regression, predicts test classes, and prints accuracy. section 8.6
