Talk
Data: crush course
Abstract
There is a fundamental difference of approach between deductive (model-based) and inductive (data-based) reasoning in sciences. In this crush course we will focus on latter approach that is rarely used in pure mathematics yet might be useful in the applied research. We cover the following models and topics:
- General principles of data-driven approach: bottle-necks and advantages;
- Regression, Classification;
- Principle Component Analysis, clustering;
- Decision trees, random forests, gradient boosting;
- Neural networks, general principles and basic models
As we cover the most general and widely used models we will come back to the first topic revisiting it of a different level.
Date and time info
Friday, 12.00-13.30
Prerequisites
Basic knowledge about real world, Analysis, general understanding of Probability and Statistics
Audience
MSc students, PhD students, Postdocs
Language
English