Practical workshop on Machine Learning in R

Malvika Sharan   2019-11-06   Comments Off on Practical workshop on Machine Learning in R

Date/Time
Date(s) - 06/11/2019 - 07/11/2019
09:00 - 17:00

Location
EMBL Heidelberg; Room 202

Categories


Tutors and helpers

– Prof Bernd Bischl
– Martin Binder
– Giuseppe Casalicchio

Affiliation: Ludwig-Maximilians-University Munich

Course Information

This two-day course on the application of Machine Learning in R will be delivered partly as an intermediate lecture on machine learning and partly as practical sessions on programming and data analysis. It introduces the new `mlr3` package, which provides a unified interface for machine learning for classification and regression. It provides many extensions for frequent use cases such as parameter tuning and preprocessing pipelines, and is being actively developed to expand its functionality. (References: homepage, manual).

Sessions will be driven by many practical exercises and case studies.

Before this workshop, participants are expected to review the official material introducing the principle of Machine Learning (see the prerequisite).

Course Content

This 2-day course will cover hands-on sessions using `mlr` and other relevant packages.

Daily schedule

– 09:30-12:30 3h morning, 90 min Theory + 90 min Practical
– 12:30-13:30 1h Lunchbreak
– 13:30-16:30 3h afternoon, 90 min Theory + 90 min Practical
– 16:30-17:00 Time for general questions

Day 1

– Lecture session: Performance Evaluation and Resampling (Metrics, CV, ROC)
– Lecture + Demo session: Introduction to mlr3 + Demo of the main functionality of mlr3 using the german credit risk data set.
– Hands-on session: Practice mlr3 using the QSAR biodegradation data set.
– Lecture session: Regularization + Boosting

Day 2

– Lecture session: Tuning + Nested CV
– Lecture + Demo session: Intro to mlr3tuning + Demo of mlr3tuning using the german credit risk data set.
– Lecture session: Feature Engineering
– Lecture + Demo session: Intro to mlr3pipelines extension + demo

Prerequisite

The course is aimed at advanced R programmers, preferably with some knowledge of statistics and data modeling (See prerequisite materials from Day-1, 2, & 4). In this course, our learners will learn more about machine learning and its application and implementation through the hands-on sessions and use cases.

Optional: Discussion-Based Session On The Principle of Machine Learning

Anna Kreshuk (EMBL Group Leader) will lead a one-day discussion-based session on 14 October 2019 to address your questions on the prerequisite materials covered in principle of Machine Learning. This will also allow you to connect with other participants of this workshop informally, and discuss the materials in smaller groups.

Please register for this workshop separately on this page: https://bio-it.embl.de/events/machine-learning-discussion-workshop-2019/.

Registration

Please note that the maximum capacity of this course is 40 participants and registration is required to secure a place. If you have any questions, please contact Malvika Sharan.

In your registration, please mention your EMBL group name, or institute’s name (e.g. DKFZ, Uni-HD) if you are registering as an external participant.

Course Fee:

All participants will be charged with a course fee of 60 Euros to support the expense for this course. The invoice details will be shared with the registered participants via email.

Internal participants can make the payment by their group’s internal budget.

Cancellation and No-Show:

The registration can be canceled for free of charge until October 31, 2019.

The participants will be charged a cancellation fee (if canceled after October 31, 2019) or no-show fee of 50 Euros.

Bookings

This event is fully booked.

Malvika Sharan

About Malvika Sharan

I am a Community Outreach Coordinator for Bio-IT and Training Coordinator for ELIXIR Germany. Find me at EMBL-HD office 101 or at Bio-IT drop-in Sessions on Tuesdays 10:00-12:00 at the EMBL staff lounge. Contact Details: email - malvika.sharan@embl.de, Twitter - https://twitter.com/MalvikaSharan, GitHub - https://github.com/malvikasharan.