mlcourse.ai is an open Machine Learning course by OpenDataScience (ods.ai), led by Yury Kashnitsky (yorko), now Staff GenAI specialist at Google Cloud. Having both a Ph.D. degree in applied math and a Kaggle Competitions Master tier, Yury aimed at designing an ML course with a perfect balance between theory and practice. Thus, the course meets you with math formulae in lectures, and a lot of practice in the form of assignments and Kaggle Inclass competitions. Currently, the course is in a self-paced mode. Here, we guide you through the self-paced mlcourse.ai.

Bonus assignments

Additionally, you can purchase a Bonus Assignments pack with the best non-demo versions of mlcourse.ai assignments. Select the "Bonus Assignments" tier on Patreon or a similar tier on Boosty (rus).

The bonus pack contains 10 assignments, in some of them you are challenged to beat a baseline in a Kaggle competition under thorough guidance ("Alice" and "Medium") or implement an algorithm from scratch -- efficient stochastic gradient descent classifier and gradient boosting.

Self-paced passing

You are guided through 10 weeks of mlcourse.ai. For each week, from Pandas to Gradient Boosting, instructions are given on which articles to read, lectures to watch, and what assignments to accomplish.

Articles

This is the list of published articles on medium.com 🇬🇧, habr.com 🇷🇺. Notebooks in Chinese 🇨🇳 are also mentioned, and links to Kaggle Notebooks (in English) are provided. Icons are clickable.

1. Exploratory Data Analysis with Pandas 🇬🇧 🇷🇺 🇨🇳, Kaggle Notebook
2. Visual Data Analysis with Python 🇬🇧 🇷🇺 🇨🇳, Kaggle Notebooks: part1, part2
3. Classification, Decision Trees and k Nearest Neighbors 🇬🇧 🇷🇺 🇨🇳, Kaggle Notebook
4. Linear Classification and Regression 🇬🇧 🇷🇺 🇨🇳, Kaggle Notebooks: part1, part2, part3, part4, part5
5. Bagging and Random Forest 🇬🇧 🇷🇺 🇨🇳, Kaggle Notebooks: part1, part2, part3
6. Feature Engineering and Feature Selection 🇬🇧 🇷🇺 🇨🇳, Kaggle Notebook
7. Unsupervised Learning: Principal Component Analysis and Clustering 🇬🇧 🇷🇺 🇨🇳, Kaggle Notebook
8. Vowpal Wabbit: Learning with Gigabytes of Data 🇬🇧 🇷🇺 🇨🇳, Kaggle Notebook
9. Time Series Analysis with Python, part 1 🇬🇧 🇷🇺 🇨🇳. Predicting future with Facebook Prophet, part 2 🇬🇧, 🇨🇳 Kaggle Notebooks: part1, part2
10. Gradient Boosting 🇬🇧 🇷🇺, 🇨🇳, Kaggle Notebook

Lectures

Video lectures are uploaded to this YouTube playlist. Introduction, video, slides

1. Exploratory data analysis with Pandas, video
2. Visualization, main plots for EDA, video
3. Decision trees: theory and practical part
4. Logistic regression: theoretical foundations, practical part (baselines in the "Alice" competition)
5. Ensembles and Random Forest – part 1. Classification metrics – part 2. Example of a business task, predicting a customer payment – part 3
6. Linear regression and regularization - theory, LASSO & Ridge, LTV prediction - practice
7. Unsupervised learning - Principal Component Analysis and Clustering
8. Stochastic Gradient Descent for classification and regression - part 1, part 2 TBA
9. Time series analysis with Python (ARIMA, Prophet) - video
10. Gradient boosting: basic ideas - part 1, key ideas behind Xgboost, LightGBM, and CatBoost + practice - part 2

Assignments

The following are demo-assignments. Additionally, within the "Bonus Assignments" tier you can get access to non-demo assignments.

1. Exploratory data analysis with Pandas, nbviewer, Kaggle Notebook, solution
2. Analyzing cardiovascular disease data, nbviewer, Kaggle Notebook, solution
3. Decision trees with a toy task and the UCI Adult dataset, nbviewer, Kaggle Notebook, solution
4. Sarcasm detection, Kaggle Notebook, solution. Linear Regression as an optimization problem, nbviewer, Kaggle Notebook
5. Logistic Regression and Random Forest in the credit scoring problem, nbviewer, Kaggle Notebook, solution
6. Exploring OLS, Lasso and Random Forest in a regression task, nbviewer, Kaggle Notebook, solution
7. Unsupervised learning, nbviewer, Kaggle Notebook, solution
8. Implementing online regressor, nbviewer, Kaggle Notebook, solution
9. Time series analysis, nbviewer, Kaggle Notebook, solution
10. Beating baseline in a competition, Kaggle Notebook

Kaggle competitions

1. Catch Me If You Can: Intruder Detection through Webpage Session Tracking. Kaggle Inclass
2. Predicting popularity of a Medium article. Kaggle Inclass
3. DotA 2 winner prediction. Kaggle Inclass

Building course material as a Jupyter Book

We are using Jupyter Book v1 to build the book.

1. Install uv for dependency management: pip install uv;
2. Run uv sync to install project dependencies, or uv lock --upgrade to update them;
3. Run uv run jb build mlcourse_ai_jupyter_book (note: this works with Jupyter Book v1) – this will take a while, the output is found in the mlcourse_ai_jupyter_book/_build folder. Namely, mlcourse_ai_jupyter_book/_build/html/index.html is what gets rendered at the mlcourse.ai main page.

This will show the path to your local file with the course material, e.g. file:///Users/nickname/Documents/mlcourse.ai/mlcourse_ai_jupyter_book/_build/html/index.html. You can open it in your browser to see the course material locally.

Citing mlcourse.ai

If you happen to cite mlcourse.ai in your work, you can use this BibTeX record:

@misc{mlcourse_ai,
    author = {Kashnitsky, Yury},
    title = {mlcourse.ai – Open Machine Learning Course},
    year = {2020},
    publisher = {GitHub},
    journal = {GitHub repository},
    howpublished = {\url{https://github.com/Yorko/mlcourse.ai}},
}