scikit-learn: your first classifier and regression model
10 steps · about 11 h 25 min · Beginner
- An Introduction to scikit-learn: Machine Learning in Python (video, Jake VanderPlas, YouTube, 1 h 30 min)
A scikit-learn core contributor explains what machine learning is and how the library is organised, before you write any code.
- Scikit-learn: A Beginner's Guide to Machine Learning in Python (article, DigitalOcean, 25 min)
A short first look at the fit pattern with a linear regression and a nearest-neighbor classifier on the iris dataset.
- Python Regression with scikit-learn (tutorial, pythonbook.io, 30 min)
Regression predicts a number, so you build one first and meet the prepare, split, train, predict, evaluate loop.
- Python Classification: Build a Classifier with scikit-learn (interactive, pythonbook.io, 35 min)
Interactive exercises with logistic regression, decision trees and k-nearest neighbors, plus metrics beyond accuracy.
- Introduction to Scikit-Learn For Beginners (article, Zero To Mastery, 25 min)
A full classifier workflow on heart-disease data that reinforces choosing a model, fitting, predicting and improving.
- Getting Started — scikit-learn documentation (doc, scikit-learn.org, 20 min)
The official overview ties together fitting, preprocessing, pipelines and why pipelines prevent data leakage.
- Introduction to Machine Learning with scikit-learn (video series and notebooks) (video, Kevin Markham, Data School (GitHub), 2 h 30 min)
Watch the first few videos of this series (it links to the YouTube playlist) and run the notebooks to practise train/test splits and model comparison.
- 3.1. Cross-validation: evaluating estimator performance (doc, scikit-learn.org, 30 min)
One train/test split can mislead you, so learn to score a model across several splits.
- Machine learning in Python with scikit-learn (course) (interactive, Inria / scikit-learn developers, 4 h)
A free, CC-BY course by the library's core developers. Work through the predictive modeling pipeline and model selection modules in depth.
- User Guide — scikit-learn documentation (reference, scikit-learn.org, 40 min)
Your reference for choosing the next model, metric or tuning method once the basics feel familiar.