Using R for Regression and Machine Learning in Investment (Coursera)

The course is designed with the assumption that most students already have a little bit of knowledge in financial economics and R programming. Students are expected to have heard about stocks and bonds and balance sheets, earnings, etc., and know the introductory statistics level, such as mean, median, distribution, regression, etc. Students are also expected to know of the instructors' 1st course, 'Fundamental of data-driven investment.'

The instructor will explain the detail of R programming. It will be an excellent course for you to improve your programming skills but you must have basic knowledge in R. If you are very good at R programming, it will provide you with an excellent opportunity to practice again with finance and investment examples.

What You Will Learn

- Understanding the basic common concept of machine learning

- Familiarizing with most commonly used methodology, regression

- Distinguishing in-sample and out-of-sample results and leading to well-performing models in a real-life

Syllabus

WEEK 1

Understanding the big picture of the algorithm-driven investment decision-making process using machine learning and review of regression methodology

Understand the characteristics of predictive models and various data in investment

The instructor will give you the big picture of the algorithm-driven investment decision-making process. After you understand that, we will review the regression concept and connect it with the core concepts of machine learning methodologies.

WEEK 2

Regression and beyond

Use regression methodology for various investment analysis purpose and improve models by using ridge, lasso, and logistic regression. First of all, you will learn how you can gauge investment strategy using backtesting.

You learned the first component of investment strategy, returns, in the first week. You will expand your study to assessing investment risks. To understand stocks' risks, you will calculate covariance and correlation matrix using historical time-series stock return data. You will extend this to market factor and three-factor models to understand the risk you are facing with your investment. Finally, you will calculate factor exposure using a 3-factor model from week 2 and separate common factor risk and idiosyncratic risk of the stock.