Evaluating Model Performance (accuracy, R²)

Model evaluation is about measuring how well a trained model makes predictions.

The metric you choose depends on the type of problem:

• Classification: accuracy, precision, recall, F1-score
• Regression: R² (coefficient of determination), MSE, MAE

Scikit-learn provides built-in functions to calculate these metrics.

The following example shows how to use accuracy score to evaluate a classification model.

Copy
import piplite
await piplite.install('scikit-learn')

from sklearn.datasets import load_iris
from sklearn.model_selection import train_test_split
from sklearn.neighbors import KNeighborsClassifier
from sklearn.metrics import accuracy_score

# Load data
X, y = load_iris(return_X_y=True)

# Split
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.3, random_state=42)

# Train classifier
knn = KNeighborsClassifier(n_neighbors=3)
knn.fit(X_train, y_train)

# Predict
y_pred = knn.predict(X_test)

# Evaluate
acc = accuracy_score(y_test, y_pred)
print(f"Accuracy: {acc:.2f}")

Accuracy = fraction of correct predictions.

Good for balanced datasets, but can be misleading if the classes are imbalanced.

The following example shows how to use R² score to evaluate a regression model.

Copy
from sklearn.linear_model import LinearRegression
from sklearn.metrics import r2_score

# Synthetic regression data
import numpy as np
rng = np.random.RandomState(0)
X_reg = 2 * rng.rand(50, 1)
y_reg = 4 + 3 * X_reg.ravel() + rng.randn(50)

# Train/test split
X_train, X_test, y_train, y_test = train_test_split(X_reg, y_reg, test_size=0.2, random_state=42)

# Train
reg = LinearRegression()
reg.fit(X_train, y_train)

# Predict
y_pred = reg.predict(X_test)

# Evaluate
r2 = r2_score(y_test, y_pred)
print(f"R² Score: {r2:.3f}")

R² measures how much variance in the target is explained by the model.

The following are the possible values of R²:

• 1.0: perfect prediction
• 0: no better than mean prediction
• Negative: worse than mean prediction

• Use classification metrics for categorical outputs, and regression metrics for continuous outputs.
• Always evaluate on a test set that the model hasn't seen during training.
• Consider multiple metrics for a more complete evaluation, especially with imbalanced data.