Feature Engineering for Improving Learning Environments (edX)

This course pays particular attention to the ways in which researchers and data scientists can transform raw data into features (i.e., variables or predictors) used in various machine learning algorithms. We will provide strategies for using prior research, knowledge from practice, and logic to create features, as well as build and evaluate machine learning models. The process of building features will be discussed within a broader data-intensive research workflow using R.

What you'll learn

- How to transform and visualize data using R

- How to apply selected machine learning algorithms (e.g., logistic regression and decision trees) to regression and classification tasks in R

- Strategies for applying data-intensive research workflows for feature engineering and model building

Syllabus

Week 1: Finding features

Introduction to setting up a feature engineering workflow, which includes identifying problems of practice, relevant research, and brainstorming potential features.

Week 2: Data wrangling and visualization

Introduction to data wrangling, data visualization techniques, and structure discovery algorithms. Integrating theory, knowledge from practice, logic, and contextual factors into feature engineering will also be discussed.

Week 3: Modeling features

Introduction to using features within selected machine learning algorithms (e.g. logistic regression and decision tree) and the tradeoffs between interpretability and prediction.

Prerequisites

We highly recommend that you take the previous course in this series before beginning this course:
Predictive Modeling in Learning Analytics

This course is intended for those who have a bachelor’s degree and are interested in developing learning and data science skills for employment in education, corporate, nonprofit, and military sectors. Experience with programming and statistics will be beneficial to participants.