Introduction to Data Science (4V, 2Ü) Prof. Ernst, WS 2023/24
Content
Tentative list of course topics:
- Introduction: What is Data Science
- Learning Theory
- Regression
- Classification
- Clustering and Tree-Based Methods
- Unsupervised Learning
- Mathematical Aspects of Neural Networks
Notices
| Note for Wednesday's class October 11, 2023 | In place of a lecture Jan Blechschmidt will hold an exercise class in the lecture room. Please bring you own laptop to class, as there will be instructions for installing a python environment locally as well as access instructions for the department's jupyterHub for the programming exercises to be done during the cour23. |
|---|---|
| First Class | Monday, October 09, 2023. |
| First Exercise Lab | Wednesday, October 11, 2023. |
Listing of this course in the electronic Vorlesungsverzeichnis (course directory):
Keine Lehrveranstaltung gefunden.
Lecture
Literature
- James, Witten, Hastie & Tibshirani. An Introduction to Statistical Learning – with Applications in Python. Springer 2023. Follow this link to access the book online.
- The authors of ISL have produced a video course based on their book which is available on YouTube.
- Here's a continually updated annotated reading list for the course (31.01.2024).
Slides
- What is Data Science? (09.10.2023)
- Learning Theory (18.10.2023)
- Linear Regression (20.11.2023)
- Classification (12.12.2023)
- Resampling Methods (20.12.2023)
- Linear Model Selection and Regularization (18.01.2024)
- Nonlinear Regression Models (22.01.2024)
- Tree-Based Methods (22.01.2024)
- Unsupervised Learning (31.01.2024)
Exercises
The exercise labs will take place in the computer pool of the Mathematics Computing Center, located in Room 738 of the building Reichenhainer Str. 39/41. To carry out the programming tasks, everybody needs a valid login which can be picked up at Margit Matt's office (Rh 39/room 704). All programming exercises can be carried out on our Jupyter Hub or on your own laptop. The Jupyter Hub can be accessed from inside the university network or via VPN.Problem sheets
The problem sheets and homeworks are available here.Installation
If you want to install your own Jupyter environment, we suggest the following procedure: Get the latest version of miniconda from this page and follow the steps in the installation dialogue. Create a new environment viaconda create --name ds2023Activate the
ds2023 environment via
conda activate ds2023or (depending on your system)
source activate ds2023and install the required packages (this list is continually updated)
conda install -c conda-forge numpy scipy pandas matplotlib scikit-learn jupyter jupyterlabPlease refer to Conda (Installation) and Conda (Managing environments) for further information.