Loved by learners at thousands of companies
Do you know the basics of supervised learning and want to use state-of-the-art models on real-world datasets? Gradient boosting is currently one of the most popular techniques for efficient modeling of tabular datasets of all sizes. XGboost is a very fast, scalable implementation of gradient boosting, with models using XGBoost regularly winning online data science competitions and being used at scale across different industries. In this course, you'll learn how to use this powerful library alongside pandas and scikit-learn to build and tune supervised learning models. You'll work with real-world datasets to solve classification and regression problems.
Classification with XGBoostFree
This chapter will introduce you to the fundamental idea behind XGBoost—boosted learners. Once you understand how XGBoost works, you'll apply it to solve a common classification problem found in industry: predicting whether a customer will stop being a customer at some point in the future.Welcome to the course!50 xpWhich of these is a classification problem?50 xpWhich of these is a binary classification problem?50 xpIntroducing XGBoost50 xpXGBoost: Fit/Predict100 xpWhat is a decision tree?50 xpDecision trees100 xpWhat is Boosting?50 xpMeasuring accuracy100 xpMeasuring AUC100 xpWhen should I use XGBoost?50 xpUsing XGBoost50 xp
Regression with XGBoost
After a brief review of supervised regression, you'll apply XGBoost to the regression task of predicting house prices in Ames, Iowa. You'll learn about the two kinds of base learners that XGboost can use as its weak learners, and review how to evaluate the quality of your regression models.Regression review50 xpWhich of these is a regression problem?50 xpObjective (loss) functions and base learners50 xpDecision trees as base learners100 xpLinear base learners100 xpEvaluating model quality100 xpRegularization and base learners in XGBoost50 xpUsing regularization in XGBoost100 xpVisualizing individual XGBoost trees100 xpVisualizing feature importances: What features are most important in my dataset100 xp
Fine-tuning your XGBoost model
This chapter will teach you how to make your XGBoost models as performant as possible. You'll learn about the variety of parameters that can be adjusted to alter the behavior of XGBoost and how to tune them efficiently so that you can supercharge the performance of your models.Why tune your model?50 xpWhen is tuning your model a bad idea?50 xpTuning the number of boosting rounds100 xpAutomated boosting round selection using early_stopping100 xpOverview of XGBoost's hyperparameters50 xpTuning eta100 xpTuning max_depth100 xpTuning colsample_bytree100 xpReview of grid search and random search50 xpGrid search with XGBoost100 xpRandom search with XGBoost100 xpLimits of grid search and random search50 xpWhen should you use grid search and random search?50 xp
Using XGBoost in pipelines
Take your XGBoost skills to the next level by incorporating your models into two end-to-end machine learning pipelines. You'll learn how to tune the most important XGBoost hyperparameters efficiently within a pipeline, and get an introduction to some more advanced preprocessing techniques.Review of pipelines using sklearn50 xpExploratory data analysis50 xpEncoding categorical columns I: LabelEncoder100 xpEncoding categorical columns II: OneHotEncoder100 xpEncoding categorical columns III: DictVectorizer100 xpPreprocessing within a pipeline100 xpIncorporating XGBoost into pipelines50 xpCross-validating your XGBoost model100 xpKidney disease case study I: Categorical Imputer100 xpKidney disease case study II: Feature Union100 xpKidney disease case study III: Full pipeline100 xpTuning XGBoost hyperparameters50 xpBringing it all together100 xpFinal Thoughts50 xp
In the following tracksMachine Learning Scientist
Head of Data Science, TelevisaUnivision
I enjoy applying my quantitative skills to large-scale data-intensive problems and mentoring junior colleagues. I am also an avid learner and am always trying to refine my programming chops and apply state of the art analytical and statistical methods. In my current role as Head of Data Science at Univision, I build proprietary data products that allow us to efficiently engage with and grow our audience. I also enjoy sharing and communicating what knowledge I have. To that end, and when time permits, I teach data science courses at several NYC area bootcamps/hacker academies/universities. Prior to Univision, I built a state of the art cross-platform media measurement solution at Viacom, automated back office processes using machine learning for clients in the financial industry, and worked at small cybersecurity and digital advertising startups. I obtained my Ph.D. in Cognitive Neuroscience at Dartmouth College.