Skip to main content

Introduction to Machine Learning

This is a fast paced 10 week course in Machine Learning for learners with background in Statistics, Linear Algebra and Probability. The course is an intermediate course in supervised learning and will focus on Model Selection and Classification techniques. This course is an interactive course where the concepts taught during lectures will be implemented during the live sessions. This course will use Python and R for all programming assignments.

About This Course

Machine learning is the science of teaching machines how to act independently by training them using historical data. With the availability of cheap computing power and digitization of data we are witnessing a paradigm shift where computer is taking a pivotal place in human life. A well programmed computer can now handle lot of seemingly complex tasks like driving cars, web search, speech recognition, sentiment analysis, sorting mails, robotics and many other applications. In this course you will learn techniques which can be used for training computers to produce reliable future forecasts based on historical data.

Course Contents

  • Data ETL, statistics and plotting
  • Linear Regression including Regularization techniques
  • Vaiables Selection - For Regressions, PCA, Correlated Variables AIC and CV use
  • Model Selection, Test/Train CV, Caret and scikit-learn. AIC, BIC, Bias Variance tradeoff, Use of test/train split and CV use
  • Classification Techniques
  • Ensemble Models: Trees, Random Forest, Gradient Boosting. Variable selection with ensemble techniques
  • Support Vector Machine
  • Practicum Analysis of a data set

Course Objectives

At the completion of the course, students will be able to do the following:
  • Select most appropriate predictive model without overfitting
  • Use variable creation and variable selection
  • Understand Bias Variance Trade off in modeling
  • Correctly estimate model parameters
  • Be able to clean a dataset and create models

Requirements

  • Comfortable with Python Programming
  • Some familiarity with R Programming or willingness to learn R on the go
  • Strong Background in Statistics and Probability (mean, variance, correlation, regression etc)
  • Good understanding of Matrix Algebra (Matrix Multiplication, Matrix Inversion)
Length: 10 Week
Effort Required: 90 minutes lecture + 5 hours per week for Projects and Assignments
Live Lectures: Wednesdays 9:00pm NYC Time(Thurdays 6:30 AM India Time)
Price: $249

Course Staff

Course Staff Image #1

Niels Nygaard

Niels O Nygaard is a Professor in the Department of Mathematics at the University of Chicago since 1982. Prior to joining the University of Chicago he taught at Princeton University. He holds a Ph.D. in Mathematics from MIT. Professor Nygaard was the Director of the Financial Mathematics Program since its inception in 1996 until April 2010. Besides his interest in Financial Mathematics he has done research in Arithmetic, Algebraic Geometry and Number Theory. Professor Nygaard was the founding director of both the Financial Mathematics Program and the Stevanovich Center for Financial Mathematics. Professor Nygaard is a recognized scientist in the field of financial mathematics and has a long list of publications

Instructor

Adam Ginensky

Dr. Ginensky has a M.S. and a Ph.D. in mathematics, both from the University of Chicago. He is currently teaching at the University of Chicago in their Master in Predictive Analytics Program. Prior to 2008, Adam worked as a market maker at the Chicago Mercantile Exchange and was involved in the mathematics of option pricing, but primarily was a floor trader. He has gave a number of talks in the University of Chicago Mathematical Finance Program. After 2008 he worked as a quantitative analyst for a proprietary trading company where he used Matlab and R (as well as SQL and various extensions) to perform data mining and statistical analysis of various financial data sets. His responsibilities included analyzing large (tick) data sets, performing statistical modeling of various time series of trading data, and writing the software packages to implement these goals. It was at this point that he became interested in applying statistics in other fields as well as finance. His current interests include both supervised and unsupervised learning as well as time series analysis. He is also currently exploring applications of algebraic geometry to statistics (algebraic statistics). In all aspects of his research and activity, he is fascinated by the practical applications of the theoretical ideas.

Enroll