DIVCLUS-T: A monothetic divisive hierarchical clustering method

Abstract

DIVCLUS-T is a divisive hierarchical clustering algorithm based on a monothetic bipartitional approach allowing the dendrogram of the hierarchy to be read as a decision tree. It is designed for either numerical or categorical data. Like the Ward agglomerative hierarchical clustering algorithm and the k-means partitioning algorithm, it is based on the minimization of the inertia criterion. However, unlike Ward and k-means, it provides a simple and natural interpretation of the clusters. The price paid by construction in terms of inertia by DIVCLUS-T for this additional interpretation is studied by applying the three algorithms on six databases from the UCI Machine Learning repository.

Fast Minimum Spanning Tree Based Clustering Algorithms on Local Neighborhood Graph

“…In an effort to maximize the distance between the centers of two partitions, BMGClust algorithm makes use of Bimeans algorithm which is a recursive bi-partitioning of the dataset with two centers chosen in a wise manner. The Bi-partitioning method recursively split the dataset into a binary tree of partitions with different partitioning criteria [14]. The Bi-means algorithm used in this paper makes use of bi-partitioning method but in a different manner.…”

Divisive Property-Based and Fuzzy Clustering for Sequence Analysis

“…The aim in using property-based clustering is to build a sequence typology by identifying well-defined clustering rules that are based on the most relevant properties of the analyzed object. In the literature, these clustering methods are called monothetic divisive clustering methods (Chavent et al 2007), and they were first introduced in sequence analysis by Piccarreta and Billari (2007). We propose here a conceptual presentation of the "DIVCLUS-T" algorithm (a detailed presentation can be found in Chavent et al 2007).…”

“…Monothetic clustering-which here we call "property-based clustering"-aims to create a sequence typology defined by explicit classification rules. In this paper, we introduce a method that is based on the "DIVCLUS-T" algorithm proposed by Chavent et al (2007). Following the work of Piccarreta and Billari (2007), we also discuss its use in sequence analysis and extend their work by proposing for consideration new sets of state sequence features.…”

DSMC: A Novel Distributed Store-Retrieve Approach of Internet Data Using MapReduce Model and Community Detection in Big Data

“…Divisive method and agglomerative method [15][16][17][18] are classical algorithms of community detection. The commonly used algorithms are GN (Girvan and Newman) algorithm and Newman fast algorithm.…”