Fuzzy Application

A Novel Approach to Noise Clustering for Outlier Detection

2009-03-31T09:45:00.000-07:00

Abstract Noise clustering, as a robust clustering method, performs partitioning of data sets educing errors caused by outliers. Noise clustering defines outliers in terms of a certain distance, which is called noise distance. The probability or membership degree of data points belonging to the noise cluster increases with their distance to regular clusters. The main purpose of noise cluster ing is to reduce the influence of outliers on the regular clusters. The emphasis is not put on exactly identifying outliers. However, in many applications outliers contain important information and their correct identification is crucial. In this paper we present a method to estimate the noise distance in noise clustering based on the preser vation of the hypervolume of the feature space. Our examples will demonstrate the efficiency of this approach.

Calculus of Fuzzy Semantic Typing for Qualitative Analysis of Text

2009-03-31T09:42:00.000-07:00

Statistical approaches to text mining can be enhanced and improved through the qualitative representation
of free text – ideally, a representation which accommodates ambiguity and imprecision. We introduce a specialized lexicon that assigns semantic categories to words, together with numeric values for centrality and intensity within each category. From this lexicon, we automatically generate an additional set of resources to implement some of the common operations of text mining – profiling, querying, and query/profile expansion and compression – in qualitative domains. We exploit the hierarchical structure of free text (i.e., sentence/ paragraph/document) and develop a set of operators whose arguments are fuzzy representations ("profiles") of text at any hierarchical level. Various operators compute the centrality and intensity of categories within a profile, a profile's overall intensity, and the cardinality and fuzziness of a profile; others are used in profile merging, profile expansion or compression, and discovery of related categories from a profile. We address the meaning and modes of deployment of these operators using practical examples. Finally, we discuss the utility of fuzzy typing for various tasks, such as "qualitative browsing" and similarity estimates. We discuss how the existing approach can be enhanced using automatic lexicon expansion and information extraction techniques. We offer a practical software demonstration with several visualization examples, illustrating the power of the proposed operators in affect analysis of news reports and movie reviews.

Fuzzy Types Clustering for Microarray Data

2009-03-31T09:40:00.000-07:00

The main goal of microarray experiments is to quantify the expression of every object on a slide as precisely as possible, with a further goal of clustering the objects. Recently, many studies have discussed clustering issues involving similar patterns of gene expression. This paper presents an application of fuzzy-type methods for clustering DNA microarray data that can be applied to typical comparisons. Clustering and analyses were performed on microarray and simulated data. The results show that fuzzy-possibility c-means clustering substantially improves the findings obtained by others.

Kernel-based fuzzy and possibilistic c-means clustering

2009-03-31T09:25:00.000-07:00

The 'kernel method' has attracted great attention with the development of support vector machine (SVM) and has been studied in a general way. In this paper, this 'method' is extended to the well-known fuzzy c-means (FCM) and possibilistic c-means (PCM) algorithms. It is realized by substitution of a kernel-induced distance metric for the original Euclidean distance, and the corresponding algorithms are called kernel fuzzy c-means (KFCM) and kernel possibilistic c-means (KPCM) algorithms. And some test results are given to illustrate the advantages of the proposed algorithms over the FCM and PCM algorithms.