Description

The course will give the student an introduction to and a basic knowledge of Machine Learning (ML) and its use in various parts of data analysis. The focus will be on application through examples and use of computers, and be project based.

The course will cover the following subjects:

• Introduction to Machine Learning
• Types of problems suitable for ML and their typical solutions.
• Types of problems not suitable for ML
• Classification and Regression
• Supervised vs. Unsupervised training
• Dimensionality Reduction
• ML performance and loss functions

Knowledge:

The student will obtain knowledge about ML concepts and procedures, more specifically:

• The fundamental methods used in ML.
• Various Loss-Functions and Goodness measures.
• The most commonly used tree and neural net based ML algorithms.
• Examples of ML usage on various types of data.
• Methods to evaluate performance and importance of individual input variables.

 

Skills:

The student should in the course obtain the following skills:

• Understand the use of ML in data analysis
• Be able to apply ML algorithms to (suitable) dataset
• Be able to optimise the performance of the ML algorithm
• Be capable of quantifying and comparing ML performances

 

Competences:

This course will provide the students with an understanding of ML methods and knowledge of (structured) data analysis with ML, which enables them to analyse data using ML in science and beyond. The students should be capable of handling data sparcity, non-uniformities, along with unbalanced and categorical data.

Recommended qualifications

Basic knowledge of programming is required corresponding to a bachelor course in programming for physicists.

The student should be familiar with the general line of thinking in programming, and be able to build own programs independently. Elementary mathematics (calculus, linear algebra, and combinatorics) is also required.



Academic qualifications equivalent to a BSc degree is recommended.