The aim of the course is to learn machine learning methods by applying them to practice-oriented exercise data sets. During the training, the central steps such as preparatory data management, training of algorithms as well as forecasting and validation are learned and directly implemented in Python. A special focus is put on the Python library scikit-learn, which includes a variety of popular algorithms in the field of machine learning.
During the course, participants will create Python scripts which can be used as templates for their own machine learning applications.
Table of contents:
- Introduction to the basic concepts of machine learning
- Dealing with the machine learning framework scikit-learn
- Introduction to machine learning algorithms such as decision trees, support vector machines or random forests
- Creation of training and test data
- Parameter tuning of the models with the help of cross-validations
- Presentation of relevant processing steps such as one-hot encoding, standardization or imputation
- Presentation of different metrics of model evaluation
- For classifications: (Balanced) Accuracy, Sensitivity, Specificity, Area under the curve
- For regressions: RMSE, MAE
- Linking of preparation and modeling steps in pipeline objects
The course is aimed at people who have already had some programming experience with Python and have a basic understanding of statistics.
—
Recommended course length: Two days