This paper presents a research on clustering an image collection using multi-visual features. The proposed method extracted a set of visual features from each image and performed multi-dimensional K-Means clustering on the whole collection. Furthermore, this work experiments on different number of visual features combination for clustering. 2, 3, 5 and 7 pair of visual features chosen from a to...

The problem of determining “the right number of clusters” in K-Means has attracted considerable interest, especially in the recent years. However, to the authors’ knowledge, no experimental results of their comparison have been reported so far. This paper intends to present some results of such a comparison involving eight cluster selection options that represent four different approaches. The ...

In this paper, we propose a new document clustering method based on the K-means method (kmeans). In our method, we allow only finite candidate vectors to be representative vectors of kmeans. We also propose a method for constructing these candidate vectors using documents that have the same word. We participated in NTCIR-4 WEB Task D (Topic Classification Task) and experimentally compared our m...

This study presents a position model for evaluating the image of tourists a destination. The evaluation is based on secondary data from 1999 through 2004, using a database composed of 20,023 respondents. Data are analyzed using the K-Means data mining method. Analytical results indicate that the destination image position (DIP) model is established, and four groups of visitor are identified. Th...

Clustering, in particular k-means clustering, is a central topic in data analysis. Clustering with Bregman divergences is a recently proposed generalization of k-means clustering which has already been widely used in applications. In this paper we analyze theoretical properties of Bregman clustering when the number of the clusters k is large. We establish quantization rates and describe the lim...

One way of representing semantics could be via a high dimensional conceptual space constructed by certain lexical semantic space models. Concepts (words), represented as a vector of other words in the semantic space, can be categorized via clustering techniques into a number of regions reflecting different contexts. The conventional clustering algorithms, e.g., K-means method, however, normally...

The issue of determining “the right number of clusters” in K-Means has attracted considerable interest, especially in the recent years. Cluster overlap appears to be a factor most affecting the clustering results. This paper proposes an experimental setting for comparison of different approaches at data generated from Gaussian clusters with the controlled parameters of betweenand within-cluster...

This paper presents a generalized version of the conventional k-means clustering algorithm [Proceedings of 5th Berkeley Symposium on Mathematical Statistics and Probability, 1, University of California Press, Berkeley, 1967, p. 281]. Not only is this new one applicable to ellipse-shaped data clusters without dead-unit problem, but also performs correct clustering without pre-assigning the exact...

In this paper, the well-known k-means algorithm for searching for a locally 12 optimal partition of the setA ⊂ R is analyzed in the case if some data points occur on the 13 border of two or more clusters. For this special case, a useful strategy by implementation 14 of the k-means algorithm is proposed. 15