- Thesis:
- Bachelor in Mathematics 50%
- Author:
- Simon Fresacher
- Title:
- Lineare Regression, Ein Vergleich zwischen Kleinste-Quadrate-Schätzer und Ridge-Schätzer
- Supervisor:
- Jan JOHANNES
- Abstract:
- Regression analysis is a method for modelling relationships between variables. This bachelorthesis first deals with the theoretical foundations of regression analysis. Subsequently, the solution of regression problems using linear algorithms is presented. For this purpose, the least squares estimator and the ridge estimator are considered.One of the two methods is better than the other when dealing with data that is high-dimensional and highly correlated. We will show that the least squares method works well as long as the number of inputs is small compared to the amount of data, and there is no strong relationship between the inputs. The ridge estimator, on the other hand, can also provide good estimates for high-dimensional problems with strong correlation.
- References:
- S. Richter. Statistisches und maschinelles Lernen, Berlin, Heidelberg: Springer, 2020.
| Statistics Group | Institute for Mathematics | Faculty of Mathematics and Computer Science | University Heidelberg |