A Greedy Clustering Algorithm Based on Interval Pattern Concepts and the Problem of Optimal Box Positioning

Abstract

We consider a clustering approach based on interval pattern concepts. Exact algorithms developed within the framework of this approach are unable to produce a solution for high-dimensional data in a reasonable time, so we propose a fast greedy algorithm which solves the problem in geometrical reformulation and shows a good rate of convergence and adequate accuracy for experimental high-dimensional data. Particularly, the algorithm provided high-quality clustering of tactile frames registered by Medical Tactile Endosurgical Complex.