Enhancing Variants of K-Means

Raghavendra Chilamakur, Rajendra Prasad Kypa, Reuben Bernard Francis

ISBN 10: 6139983800 / ISBN 13: 9786139983803

Published by LAP LAMBERT Academic Publishing, 2019

About this Item

Zustand: Hervorragend | Sprache: Englisch | Produktart: Bücher | Clustering analysis is one of the most commonly used data processing algorithms. Over half a century, K-means remains the most popular clustering algorithm because of its simplicity. Traditional K-means clustering tries to assign n data objects to k clusters starting with random initial centers. However, most of the k- means variants tend to compute distance of each data point to each cluster centroid for every iteration. We propose a fast heuristic to overcome this bottleneck with only marginal increase in Mean Squared Error (MSE). We observe that across all iterations of K-means, a data point changes its membership only among a small subset of clusters. Our heuristic predicts such clusters for each data point by looking at nearby clusters after the first iteration of k-means. We augment well-known variants of k- means like Enhanced K-means and K-means with Triangle Inequality using our heuristic to demonstrate its effectiveness. For various datasets, our heuristic achieves speed-up of up-to 3 times when compared to efficient variants of k-means.

Seller Inventory # 33557199/1

Bibliographic Details

• Title: Enhancing Variants of K-Means
• Publisher: LAP LAMBERT Academic Publishing
• Publication Date: 2019
• Language: English
• ISBN 10: 6139983800
• ISBN 13: 9786139983803
• Binding: Soft cover
• Condition: Hervorragend
• Seller catalogs:
  • Bücher

About this title

Synopsis

Clustering analysis is one of the most commonly used data processing algorithms. Over half a century, K-means remains the most popular clustering algorithm because of its simplicity. Traditional K-means clustering tries to assign n data objects to k clusters starting with random initial centers. However, most of the k- means variants tend to compute distance of each data point to each cluster centroid for every iteration. We propose a fast heuristic to overcome this bottleneck with only marginal increase in Mean Squared Error (MSE). We observe that across all iterations of K-means, a data point changes its membership only among a small subset of clusters. Our heuristic predicts such clusters for each data point by looking at nearby clusters after the first iteration of k-means. We augment well-known variants of k- means like Enhanced K-means and K-means with Triangle Inequality using our heuristic to demonstrate its effectiveness. For various datasets, our heuristic achieves speed-up of up-to 3 times when compared to efficient variants of k-means.