mlcourse.ai fall2019 Live Session 0
0 likes1,564 views
mlcourse.ai is one of the largest free machine learning courses run by OpenDataScience, with over 10,000 participants and 3,000 alumni. The course consists of 4 modules covering topics like data analysis, tree-based models, linear models, and other techniques. Each module includes quizzes, assignments applying the concepts, and practice competing in Kaggle competitions to beat baselines.
mlcourse.ai fall2019 Live Session 0
- 1. mlcourse.ai by OpenDataScience lead by Yury Kashnitsky (@yorko)
- 2. In 4 lines • One of the largest open & free ML courses • Run for ~3 years by ODS.ai team • >10k participants, ~3k alumni • Good start for a junior DS position A podcast about mlcourse.ai
- 3. Fall 2019 design choice • 4 modules • Quizzes (theory) • Assignments (practice) • Kaggle (practice)
- 4. Module 1. Data Analysis The golden rule: in your task, always start with Exploratory Data Analysis. Plan: • 1 assignment • Dota 2 winner prediction EDA - Kaggle
- 5. Module 2. Tree-based models Random Forest and gradient boosting are wheelhorses of practical ML Plan: • 1 assignment in 2 parts • 1 quiz • Beating a baseline in a Kaggle competition
- 6. Module 3. Linear models Linear models are most wide-spread in whole of ML, econometrics, statistics etc. Plan: • 1 quiz • Beating a baseline in a Kaggle competition + actually competing
- 7. Module 4. “The rest” How to scale to Gbs with simple models (Vowpal Wabbit), time series, unsupervised learning Plan: • 1 quiz (all topics) • 1 assignment (time series) • Beating a baseline in a Kaggle competition with ~10 Gb of data
- 8. Roadmap What you can follow: • #mlcourse_ai_news • mlcourse.ai/roadmap • Google calendar