Comparisons between Automatic and Non-Automatic Clustering Algorithms

Abstract

This paper presents a comparative study between two famous types of clustering algorithms. These types are the automatic and non-automatic clustering algorithms. The comparisons concerned some different criteria such as: dataset size, clusters number, execution time, results quality and accuracy. An effective automatic clustering algorithm is chosen as a sample for the automatic clustering techniques, while the well-known partitional K-Means clustering algorithm is taken as a sample for the non-automatic clustering techniques. The two chosen algorithms are implemented on the same database (ORL) concerning the human face images. Some conclusions are extracted to the performance of this implementation. MATLAB version (R2010a) is used to achieve the purpose of this paper.