Machine Learning can help to identify patterns in complex data sets, understand factors impacting a specific outcome, and increase the quality of data-driven decisions through more accurate predictions. This course focuses on supervised and unsupervised machine learning techniques for the analysis and evaluation of data. Besides covering different types of statistical machine learning models needed for complex (business) analytics problems, it also considers aspects of data preparation, model development, training, evaluation, and model parameter optimization.

On successful completion of the course, you should be able to:

• Demonstrate in-depth knowledge of supervised and unsupervised machine learning models, specifically for Regression, Classification, and Clustering
• Train, evaluate, and apply supervised and unsupervised learning models on business analytics problems
• Correctly interpret results of machine learning models and make appropriate (business) recommendations
• Demonstrate an understanding of the requirements and assumptions of the various models
• Implement techniques for data preparation, model selection, and model performance improvements
• Demonstrate knowledge of the important parameters of the considered methods and an ability to optimize the parameters for a given dataset
• Independently conduct a machine learning project using a state-of-the-art statistical tool

Students are expected to attend all meetings of this course.

The course will cover the entire lifecycle of a machine learning project, with specific emphasis on the training, evaluation, and optimization of different types of machine learning models. It will use a combination of lectures and hands-on implementation sessions to introduce, analyze, and evaluate different machine learning concepts. As part of the course, students will work on a semester-long course project in which they apply the various machine learning techniques learned in class to a data set of their choice (e.g., from their own research). The results of the course project will be written up in a research paper and presented at the end of the course.

The assessment of the course is based on following components:

• Course project, milestone 1: 40%. This includes the description of the scenario and data set as well as data preparation activities
• Course project, milestone 2: 40%: The final, written report including all analyses, evaluations, and description of results.
• Project presentation: 20%. The presentation of the course project results. 

Grading scheme

• Excellent ≥87,5%
• Good ≥75,0%
• Satisfactory ≥62,5%
• Sufficient ≥50,0%
• Fail <50,0%

The course builds on prior statistics/mathematics courses and targets PhD students/ who want to learn more about machine learning techniques and apply it to their own research. As this is not an introductory statistics course, students are expected to have taken prior introductory statistics / mathematics courses covering basic statistical methods, particularly (linear) regression.

Christian Haas