Evaluating Machine Learning Model Quality

Introduction

In machine learning, it is important to evaluate the quality of the predictions made by a model. This helps us understand how well the model is performing and whether it can be trusted for making accurate predictions. The scikit-learn library provides several metrics and scoring methods to quantify the quality of predictions.

In this lab, we will explore three different APIs provided by scikit-learn for model evaluation: the Estimator score method, the scoring parameter, and the metric functions.

VM Tips

After the VM startup is done, click the top left corner to switch to the Notebook tab to access Jupyter Notebook for practice.

Sometimes, you may need to wait a few seconds for Jupyter Notebook to finish loading. The validation of operations cannot be automated because of limitations in Jupyter Notebook.

If you face issues during learning, feel free to ask Labby. Provide feedback after the session, and we will promptly resolve the problem for you.

Skills Graph

Estimator Score Method

The Estimator score method is a default evaluation criterion provided by scikit-learn for each estimator. It calculates a score that represents the quality of the model's predictions. You can find more information about this in the documentation of each estimator.

Here's an example of using the score method for an estimator:

from sklearn.linear_model import LogisticRegression
from sklearn.datasets import load_digits

X, y = load_digits(return_X_y=True)
clf = LogisticRegression()
clf.fit(X, y)

score = clf.score(X, y)
print("Score:", score)

Scoring Parameter

Scikit-learn provides a scoring parameter in several model evaluation tools, such as cross-validation and grid search. The scoring parameter controls the metric applied to the estimators during evaluation.

Here's an example of using the scoring parameter with cross-validation:

from sklearn.model_selection import cross_val_score
from sklearn.linear_model import LogisticRegression
from sklearn.datasets import load_digits

X, y = load_digits(return_X_y=True)
clf = LogisticRegression()

scores = cross_val_score(clf, X, y, cv=5, scoring='accuracy')
print("Scores:", scores)

Metric Functions

The scikit-learn metrics module implements several functions for assessing prediction error for specific purposes. These functions can be used to calculate the quality of predictions made by a model.

Here's an example of using the accuracy_score function from the metrics module:

from sklearn.metrics import accuracy_score
from sklearn.linear_model import LogisticRegression
from sklearn.datasets import load_digits
from sklearn.model_selection import train_test_split

X, y = load_digits(return_X_y=True)
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2)

clf = LogisticRegression()
clf.fit(X_train, y_train)

y_pred = clf.predict(X_test)
accuracy = accuracy_score(y_test, y_pred)
print("Accuracy:", accuracy)

Summary

In this lab, we learned about three different APIs provided by scikit-learn for model evaluation: the Estimator score method, the scoring parameter, and the metric functions. These APIs allow us to evaluate the quality of predictions made by a model and understand how well the model is performing.