Machine Learning and Optimization
Start Date: February 27, 2019
End Date: April 17, 2019
Machine Learning and Optimization is one of five non-credit courses in the Certification in Practice of Data Analytics (CPDA) program. The course is taught by faculty from the College of Engineering at The Ohio State University. The course is delivered in 100% distance learning format and includes instructional material equivalent a one semester credit hour class.
This course can be taken individually, or as one of four courses required to receive the CPDA certificate of completion. It is expected that participants will have taken Foundations of Statistics first and Data Mining before this course and Linear Algebra and Calculus for Machine Learning if necessary.
This course will develop a solid background for understanding the theory of machine learning and applying it in real world setting. The topics covered during the course include the history of machine learning, supervised and unsupervised learning methods, linear and logistic regression, classification problems, support vector machines (SVM), neural network, and deep learning. Each class will begin with a concrete real-world learning problem, which we will then formulate mathematically, and then develop an algorithm to solve the problem. The course will use R extensively for completing all the assignments. Towards the end of the class, we will also discuss other modern solution approaches for solving machine learning problems. The evaluation will consist of six programming assignments.
4 CEUs are granted upon successful completion of the course.
You Will Learn to:
- Understand algorithms to solve unconstrained and constrained optimization problems.
- Learn basic supervised and unsupervised learning algorithms and their mathematical underpinnings.
- Evaluate which learning algorithms are useful for what kind of tasks.
- Assess the limits of machine learning.
Familiarity with calculus, linear algebra (matrices, determinants, eigenvalues, and eigenvectors) convex functions and sets, differentiation of multivariate functions, taylor series for multivariate functions, and basic probability and statistics (random variables, expectation, mean and covariance, characteristic functions, central limit theorem, etc.) If you're pursuing the CPDA certification, Foundation of Statistics must be taken first and also Data Mining before Machine Learning.
This class requires you to use the statistical software package called R (The R Project for Statistical Computing;(http://www.r-project.org/). This software package is available as Free Software.
- From the CRAN archive at https://cran.r-project.org, you can download R for Windows, Mac, and Linux.
- An in-depth introduction to R is available at http://cran.r-project.org/doc/manuals/R-intro.pdf
- Hands-on tutorials are available in the Swirl system, which you can learn about at http://swirlstats.com/. In particular, “R Programming: The basics of programming in R” is an appropriate first tutorial for students who have never used R.
- An easier to use interface to R is available in the software package RStudio. This package is available for Windows, Mac, and Linux and can be downloaded for free from http://rstudio.org.
Are You Prepared For This Course?
Below is a link to a quiz you can use to assess your background and preparedness for this course. If you are not certain if you meet the pre-req's, or if it has been some time since you took coursework in linear algebra and calculus, this quiz will help. It is not graded. It's a self-assessment with instructions included by the course instructor. If you fail or struggle with the quiz, it's recommended that you complete the Linear Algebra and Calculus for Machine Learning course before taking Machine Learning.
Assessment Quiz for Machine Learning
Click Here to learn more about how this course is delivered 100% online!
Expected Time Commitment to Complete this Course
Each course is equivalent to a one semester credit hour class. Therefore each class consists of approximately 40 hours of class time that includes 12-13 hours of recorded faculty lectures and 23-24 hours of additional course work. Each course is seven weeks in length, so each week there is 5.7 hours of combined class time (40 hrs / 7 weeks). The average student should allow a 2:1 study-to-class-time ratio to complete the course. This means you should plan to study two hours for each one hour of class time. This equates to 11-12 hours each week to complete all course work. (5.7 hrs X 2 = 11-12 hrs). Based on a person's own personal strengths and experience, you should increase or decrease the ratio.
Cancellations and Refunds
A full refund minus a $50 administrative fee will be made if cancellation is received three weeks prior to the start of the course. No refunds within three weeks of the course start date.
Course Offering Dates
Each course offering in this program is faculty lead and therefore operates with a specific start date and end date. Students must complete each course during the specific time frame. Access to the online course and materials is removed when the course ends.