Data fitting is all around us – from engineering applications, finance, and medicine to machine learning. With the growing amount of available data, it becomes an increasingly challenging computational task. This is where numerical analysis comes in, constantly seeking new algorithms and methods to tackle such problems.
Many practical problems in data fitting can be formulated as least squares problems – a classical and mathematically well-studied area. Yet, solving these problems for large-scale data remains an active research topic, and implementing efficient algorithms is far from trivial.
For verifying and comparing methods, having high-quality test data is crucial. And this is where you come in! In the world of numerical analysis, publicly available datasets are rarely used – terminology varies widely, so we often simply don’t know about these data, and they are frequently stored in unusual formats, accessible only through specific software, and so on.
What to expect during the internship?
- Explore the “terminology jungle” and identify applications corresponding to least squares problems.
- Download publicly available data and convert it into a uniform, easy-to-use format.
- Prepare data for software most commonly used by numerical analysts: Python, MATLAB, Julia…
Who we are looking for?
A curious person who enjoys discovering new things and has at least basic programming experience (e.g., in Python).
What you will gain?
The goal of the internship is to create a publicly available dataset package that can serve as a reference for other researchers – and something you can proudly showcase.
Join us and learn more about datta fitting!