Spectral clustering and investigation of the effect of different distance functions in spectral clustering

In this thesis, spectral clustering and the effect of different distance functions in spectral clustering will be examined. Instead of the Euclidean metric used in spectral clustering, it will be examined on sample data sets whether a more successful clustering can be obtained by considering different distance functions. In the first chapter, the history of spectral clustering will be mentioned and an introduction to spectral clustering will be made. In the second part, similarity graphs will be explained. In the third chapter, Laplacian matrices will be explained. In the fourth chapter, graph partitioning will be examined. In the fifth chapter, spectral clustering algorithms will be explained. In the sixth chapter, different distance functions will be handled with Python on 3 data sets and an application will be made.