Research and Application of High Dimensional Discrete Data Clustering Algorithm Based on Kernel Function
Download as PDF
DOI: 10.25236/iceeecs.2018.004
Corresponding Author
Fulan Ye
Abstract
Aiming at the disadvantages of traditional fuzzy kernel clustering algorithm, which do not consider the different contribution degree of each dimension feature to clustering, and easily fall into local optimum, an improved fuzzy kernel clustering algorithm is proposed. The algorithm constructs a simple and effective fitness function that combines the advantages of the global search of the genetic algorithm to avoid the algorithm falling into a local optimum. A weight coefficient was also introduced for each dimension feature and weighted with the Relief algorithm. This algorithm is much better than the traditional fuzzy kernel clustering algorithm. The experimental results show its effectiveness.
Keywords
Research, Application, High Dimensional Discrete Data, Clustering Algorithm, Kernel Function.