- Thesis:
- Bachelor in Mathematics
- Author:
- Michael Sucker
- Title:
- Rethinking statistical learning theory
- Supervisor:
- Jan JOHANNES
- Abstract:
- In this thesis, the aim is to develop an algorithm for a classification problem. This algorithm should create a classification rule based on given data. The work is basically divided into a theoretical part and a practical part. The theoretical part includes the derivation of the algorithm. This is done in several steps. The first of them deals with the derivation of the basic idea for the classification rule, based on a model of data generation. The following chapters then deal with the problems that arise and how to solve them. In particular, the individual chapters build on each other. Simplifying assumptions and resulting problems, which are necessary to solve a subproblem in one chapter, are always dealt with in the following one. In the end, this generates a solution that can be used in practice. The practical part deals with the implementation of the found algorithm and its evaluation. At first some basic properties are worked out by using two artificially created data sets. Afterwards, the results are demonstrated using a standard example.
References:- V.N. Vapnik und R. Izmailov. Rethinking statistical learning theory: Learning using statistical invariants. Machine Learning, 108, 2018.
| Statistics Group | Institute for Mathematics | Faculty of Mathematics and Computer Science | University Heidelberg |