Machine Learning and Data Mining with R (23.-25.03.2021)

Date: Tuesday - Thursday, 23.-25.03.2021 from 09:00-16:00 o'clock
Place: online
Language: english
Trainer: Guiseppe Casaliccio from artaro GmbH in cooperation with Essential Data Science Training
Targen group: For a maximum of 12 doctoral students who are members of the Faculty Graduate Center for Mechanical Engineering. The registration is binding. Please cancel up to 3 weeks before the start of the course. In the event of late cancellation or no-show, an apology signed by the first supervisor is required. Members of other graduate centers will be deregistered on DocGS without notification. If there are free course places after the registration deadline, members of other GZs can register by e-mail (fgz@mw.tum.de).

In this course, algorithms and general concepts of supervised machine learning are presented. Machine learning algorithms are particularly suitable for modelling non-linear relationships for complex classification and regression problems. The basic principles of the presented algorithms and concepts are explained from a theoretical point of view, their mode of operation is illustrated using the R package mlr3 and the advantages and disadvantages are discussed. All introduced algorithms and topics will be illustrated by practical examples and use cases, and are afterwards practiced by participants using the R package mlr3 based on short hands-on exercises.

Content (check list):

• Basic concepts of supervised machine learning.
• Introduction and overview of some important supervised machine learning algorithms.
• Overview of important evaluation metrics for regression and classification.
• Resampling methods to assess the performance of ML algorithms.
• Hyperparameter optimization and pitfalls when optimizing ML algorithms.
• Introduction to the machine learning R packages mlr3 and mlr3tuning with use cases and hands-on exercises.

Participation Criteria & Registration

Knowledge:

• Participants should have very good knowledge in R, i.e., they should already have worked with R and be perfectly able to perform basic tasks such as importing data, modifying data, subsetting data, as well as visualizing data using R.
• Knowledge in fitting statistical models to data using R (e.g., the linear regression model) as well as theoretical knowledge in mathematical and statistical foundations will be extremely important and advantageous in order to follow the course.

Technical requirements:

Use a laptop/PC with reliable Internet access and install the following software:

• R (at least version 3.2.0: https://cran.r-project.org)
• RStudio (at least version 1.1.X: https://www.rstudio.com/products/rstudio)
• Zoom https://zoom.us/download

Make sure that you have sufficient permissions on your laptop to install extension R packages (e.g., with the command install.packages).