Python - Python Machine Learning - Standard Deviation Part 5: Standard Deviation in Model Evaluation
In machine learning, the standard deviation is often used in cross-validation to measure model stability.
Example 6: Standard Deviation in Model Accuracy
import numpy as np
accuracies = [82, 85, 80, 78, 88] # Accuracy from different runs
mean_accuracy = np.mean(accuracies)
std_dev = np.std(accuracies)
print("Mean Accuracy:", mean_accuracy)
print("Standard Deviation:", std_dev)
Output:
Mean Accuracy: 82.6
Standard Deviation: 3.712
Explanation:
A low standard deviation (e.g., <5) means the model is stable.
A high deviation indicates inconsistent performance across different runs.