Friday, 8th May 2020, 9h-17h
postponed to Monday, 7th September 2020, 9h-17h
This training is designed for anyone who wants to understand the basic concepts of machine Learning.
This lecture provides an introduction to data analytics and machine learning, including applications in the easy-to-use software “KNIME”. Data analytics aims at analyzing data to detect informative patterns and draw conclusions about specific processes (like customer behavior, production, turnover…). One subfield is (supervised) machine learning, which aims at predicting an outcome in the future, e.g. sales, based on analyzing patterns of potentially relevant factors (or “predictors”) like price, quality, etc. in past data. This lecture discusses important tools for prediction (such as regression and decision trees), focusing on the intuition of the various methods rather than technical details. The tools are applied to the analysis of commercial data in “KNIME”, which is an open source software that is free of charge. “KNIME” is based on a graphical interface that represents all analytical steps by an intuitive flow chart and does not require programming skills. “KNIME” is therefore ideal for those interested in applying data analytics and machine learning without learning a programming language.
The lecture is given in English by Prof. Martin Huber.
- To understand the idea and goals of data analytics and machine learning
- To understand the intuition, advantages, and disadvantages of alternative methods
- To be able to apply the methods to real-world data using the software “KNIME”
- Introduction to the concept and purpose of data analytics and machine learning
- Linear and non-linear regression (OLS, logit regression)
- Penalized regression for variable selection and shrinking (lasso and ridge regression)
- Tree-based approaches (trees, bagging, random forests)
- Model tuning (cross-validation)
- Performance evaluation (out-of-sample testing)
- Application of all methods to commercial data using the statistical software “KNIME”
The lecture slides are based on “An Introduction to Statistical Learning with Applications in R” by Gareth James, Daniela Witten, Trevor Hastie, and Robert Tibshirani (Springer, New York, 2013). The text book is available as pdf at http://www-bcf.usc.edu/~gareth/ISL/.
- Participants are requested to bring their own laptop and install Knime prior to the lecture, please see the following installation guide: https://www.youtube.com/watch?v=yeHblDxakLk. (Please contact us if this is not feasible.)
- The maximum number of participants is 18.
- the participation fee is 500 CHF / 400 CHF for Swiss Engineering-section Fribourg members