A conceptual version of the K-means algorithm
Pattern Recognition Letters
Conceptual Clustering, Categorization, and Polymorphy
Machine Learning
Structuralization of universes
Fuzzy Sets and Systems
Hi-index | 0.00 |
An algorithm for automated construction of conceptual classification is presented. The algorithm arranges the objects into clusters using clustering criterion based on graph theory and constructs the concepts based on attributes that distinguish objects in the different computed clusters (typical testors). The algorithm may be applied in problems where the objects can be described simultaneously by qualitative and quantitative attributes (mixed data); with incomplete descriptions (missing data) and the number of clusters to be formed is not known a priori. LC has been tested on data from standard public databases.