Syllabus

Title
1683 Exploiting Data: The Machine Learning Approach
Instructors
Stefan Vamosi, MSc., Mgr. Jan Valendin, Dipl.-Ing.(FH) Ralf Vamosi
Type
PI
Weekly hours
2
Language of instruction
Englisch
Registration
09/17/20 to 09/24/20
Registration via LPIS
Notes to the course
Dates
Day Date Time Room
Tuesday 10/13/20 04:00 PM - 07:00 PM Online-Einheit
Tuesday 10/20/20 04:00 PM - 07:00 PM Online-Einheit
Tuesday 10/27/20 04:00 PM - 07:00 PM Online-Einheit
Tuesday 11/03/20 04:00 PM - 07:00 PM Online-Einheit
Tuesday 11/10/20 04:00 PM - 07:00 PM Online-Einheit
Tuesday 11/17/20 04:00 PM - 07:00 PM Online-Einheit
Tuesday 12/01/20 04:00 PM - 07:00 PM Online-Einheit
Tuesday 12/15/20 04:00 PM - 07:00 PM Online-Einheit
Procedure for the course when limited activity on campus

If the Covid19 measures should be reintroduced, the course will take place live via the video conferencing system MS Teams. The course dates and times would remain as scheduled. Also the homework assignments and the student's presentations would be proceeded. However, in this scenario the exam would be realized as an oral exam via MS Teams.

Contents

In customer and market analysis, the increasing amount of data available creates an immense opportunity for machine learning. This course introduces a small starter-set of simple (yet powerful) machine learning methods including their theoretical background, with practical hands-on exercises. The R programming language is used for the exercises and homeworks: prior programming experience is not required, but is recommended for successful passing of this course (see details below).

keywords: supervised and unsupervised learning, model performance evaluation, k-nearest neighbors (kNN), decision trees, random forest, k-means clustering, neural networks, support vector machines (SVM)

Learning outcomes

Our goal is to give you a set of tools you can use to make data-driven decisions wherever data is available, and hopefully inspire your future work in this exciting, dynamic field. Additionally, we aim to give an insight into some of the many modern applications of machine learning which are increasingly invading our private and professional lives. 

  • Understand the theory behind a few basic machine learning methods.
  • Get familiar with machine learning in R using real life examples.
  • Learn to chose the right method for a given task and evaluate results.
Attendance requirements

You need to attend at least 80% of all classes to pass the course.

Teaching/learning method(s)

A typical class contains a lecture and practice part, during which we guide the students through a pre-prepared programming exercise. Students are asked to prepare homeworks where methods from class are applied onto a new problem. Group challenge is a "kaggle"-style competition in which small independent groups of students compete for some bonus points. 

 

Assessment

The final grade will be evaluated as follows:

  • Class participation: [weight: 10%]
  • Homeworks (4 out of 5): [weight: 4x10%=40%]
  • Final exam: [weight: 50%]
  • Group challenge: [5% bonus]

We ask a couple of random students to present the findings from the homework at the beginning of next lecture. Doing the homework is a good start to pass the final exam, which contains open ended and multiple choice questions, with topics covering the material from the lectures. The final focuses on understanding the methods and concepts from a theoretical and practical perspective.

To pass this course: 

  • your weighted final grade needs to exceed 60%
  • you need to submit at least 3 out of 5 homeworks
  • you need at least 50% on the Final test

Grading-scheme:

< 60%                                fail (5)
60% bis 69,99%               sufficient (4)
70% bis 79,99%               satisfactory (3)
80% bis 89,99%               good (2)
>= 90%                              excellent (1)
Readings
1 Author: Brett Lantz
Title:

Machine Learning with R


Publisher: PACKT
Edition: 2nd
Recommendation: Reference literature
Type: Book
Prerequisites for participation and waiting lists

The class does not take place in a computer room.

Please bring your own laptop with a working RStudio installation. 

The software is free, runs on Windows, Mac, and Linux.

Download RStudio Desktop from: https://rstudio.com/products/rstudio/ 

Recommended previous knowledge and skills

Prior programming experience is highly recommended. For those who are not familiar with R, there are many online resources that provide the basic skills required to take this course, and we strongly encourage you to do so. The following is just a small sample of the vast resources available for free:

https://www.udemy.com/course/r-basics/

https://www.coursera.org/learn/r-programming

https://www.edx.org/course/data-science-r-basics

https://www.datacamp.com/courses/free-introduction-to-r

Availability of lecturer(s)

Consultations after prior arrangement via email.

Last edited: 2020-09-10



Back