Internship
Student login
← Data Science learning path
Data Science · Module 7 of 10

Machine learning for data science

Build predictive models with scikit-learn.

📖 Key ideas

  1. Supervised: regression (numbers) and classification (categories).
  2. Unsupervised: clustering and dimensionality reduction.
  3. Workflow: split, train, evaluate, tune.
  4. Popular models: linear/logistic regression, random forest, XGBoost, k-means.
  5. Evaluate with RMSE, accuracy, precision, recall, F1, AUC.
💡 Real-life example: Machine learning is teaching the computer with past examples so it can predict new ones.
Train/test splitSeparate data for evaluation
Random forestEnsemble of trees
Cross-validationRotating validation folds
Feature importanceWhich inputs matter most