TAMPERE UNIVERSITY OF TECHNOLOGY Master’s Degree Program in Information Technology Ranganathan, Sindhuja: Improvements to k-means clustering Master’s thesis, 42 November 2013 Major Subject: Software Systems Examiner(s): Professor Tapio Elomaa

The -means algorithm is by far the most widely used method for discovering clusters in data. We show how to accelerate it dramatically, while still always computing exactly the same result as the standard algorithm. The accelerated algorithm avoids unnecessary distance calculations by applying the triangle inequality in two different ways, and by keeping track of lower and upper bounds for dist...

This paper addresses the construction of a short-vector (128D) image representation for large-scale image and particular object retrieval. In particular, the method of joint dimensionality reduction of multiple vocabularies is considered. We study a variety of vocabulary generation techniques: different k-means initializations, different descriptor transformations, different measurement regions...

In a standard cluster analysis, such as k-means, in addition to clusters locations and distances between them, it’s important to know if they are connected or well separated from each other. The main focus of this paper is discovering the relations between the resulting clusters. We propose a new method which is based on pairwise overlapping k-means clustering, that in addition to means of clus...

Beekeeping plays an important role in increasing and diversifying the incomes of many rural communities in Kingdom of Saudi Arabia. However, despite the region’s relatively good rainfall, which result in better forage conditions, bees and beekeepers are greatly affected by seasonal shortages of bee forage. Because of these shortages, beekeepers must continually move their colonies in search of ...

Learning the number of clusters is a key problem in data clustering. We present dip-means, a novel robust incremental method to learn the number of data clusters that can be used as a wrapper around any iterative clustering algorithm of k-means family. In contrast to many popular methods which make assumptions about the underlying cluster distributions, dip-means only assumes a fundamental clus...

Traditional clustering ensembles methods combine all obtained clustering results at hand. However, we observe that it can often achieve a better clustering solution if only part of all available clustering results are combined. This paper proposes a novel clustering ensembles method, termed as resampling-based selective clustering ensembles method. The proposed selective clustering ensembles me...

This paper shows that the k-means quantization of a signal can be interpreted both as a crisp indicator function and as a fuzzy membership assignment describing fuzzy clusters and fuzzy boundaries. Combined crisp and fuzzy indicator functions are defined here as natural generalizations of the ordinary crisp and fuzzy indicator functions, respectively. An application to iris segmentation is pres...