In the Congo Basin, the elevated vulnerability of food security and the water supply implies that sustainable development strategies must incorporate the effects of climate change on hydrological regimes. However, the lack of observational hydro-climatic data over the past decades strongly limits the number of studies investigating the effects of climate change in the Congo Basin. We present th...

A new algorithm is proposed which accelerates the mini-batch k-means algorithm of Sculley (2010) by using the distance bounding approach of Elkan (2003). We argue that, when incorporating distance bounds into a mini-batch algorithm, already used data should preferentially be reused. To this end we propose using nested mini-batches, whereby data in a mini-batch at iteration t is automatically re...

A modified method for better superpixel generation based on simple linear iterative clustering (SLIC) is presented and named BSLIC in this paper. By initializing cluster centers in hexagon distribution and performing k-means clustering in a limited region, the generated superpixels are shaped into regular and compact hexagons. The additional cluster centers are initialized as edge pixels to imp...

This paper attempts to discover the most desired attributes of a model ERP implementation project. The analysis, based on research conducted among a few dozen companies implementing an ERP system, employs the statistical methods of element grouping. First, using an agglomeration method, the number of groups was determined, which was used as a parameter in the subsequently applied k-Means method...

There is often a large disparity between the size of a game we wish to solve and the size of the largest instances solvable by the best algorithms; for example, a popular variant of poker has about 10 nodes in its game tree, while the currently best approximate equilibrium-finding algorithms scale to games with around 10 nodes. In order to approximate equilibrium strategies in these games, the ...

This paper introduces a straightforward generalization of the well-known LVQ1 algorithm for nearest neighbour classifiers that includes the standard LVQ1 and the k-means algorithms as special cases. It is based on a regularizing parameter that monotonically decreases the upper bound of the training classification error towards a minimum. Experiments using 10 real data sets show the utility of t...

In this paper we give a first set of communication lower bounds for distributed clustering problems, in particular, for k-center, k-median and k-means. When the input is distributed across a large number of machines and the number of clusters k is small, our lower bounds match the current best upper bounds up to a logarithmic factor. We have designed a new composition framework in our proofs fo...

We propose a new clustering algorithm based upon the maximin correlation analysis (MCA), a learning technique that can minimize the maximum misclassification risk. The proposed algorithm resembles conventional partition clustering algorithms such as k-means in that data objects are partitioned into k disjoint partitions. On the other hand, the proposed approach is unique in that an MCA-based ap...

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