Hands-on Data Science

Catalog description

Successful machine learning systems require more than fitting complex models on data. You need to understand your dataset and all the decisions that go into data collection and modeling, carefully validate your results, and you need to be able to explain and defend the conclusions of the model. You will learn all these aspects and more in this course. Topics include exploratory data analysis, data splitting and preprocessing, cross validation and the bias-variance trade-off, training linear and nonlinear supervised machine learning models, measuring uncertainties, and making the model predictions interpretable/explainable. You will also learn about techniques to handle missing values. We use the Python data science ecosystem (e.g., sklearn, pandas and polars, matplotlib, XGBoost, SHAP). Prerequisites: Python coding experience (this course is not suitable for students with no prior python experience). Familiarity with linear algebra and calculus.