Clustering of Concept Drift Categorical Data using POur-NIR Method

Categorical data clustering is an interesting challenge for researchers in the data mining and machine learning, because of many practical aspects associated with efficient processing and concepts are often not stable but change with time. Typical examples of this are weather prediction rules and cu...

Full description

Saved in:
Bibliographic Details
Published in:International journal of computer science and information security 2011-07, Vol.9 (7), p.109-109
Main Authors: Reddy, N Sudhakar, Sunitha, K V N
Format: Article
Language:English
Subjects:
Algorithms
Clustering
Clusters
Computer information security
Data mining
Data points
Labeling
Mathematical models
Sampling
Sampling techniques
Search engines
Studies
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Categorical data clustering is an interesting challenge for researchers in the data mining and machine learning, because of many practical aspects associated with efficient processing and concepts are often not stable but change with time. Typical examples of this are weather prediction rules and customer's preferences, intrusion detection in a network traffic stream . Another example is the case of text data points, such as that occurring in Twitter/search engines. In this regard the sampling is an important technique to improve the efficiency of clustering. However, with sampling applied, those sampled points that are not having their labels after the normal process. Even though there is straight forward method for numerical domain and categorical data. But still it has a problem that is how to allocate those unlabeled data points into appropriate clusters in efficient manner. In this paper the concept-drift phenomenon is studied, and we first propose an adaptive threshold for outlier detection, which is a playing vital role detection of cluster. Second, we propose a probabilistic approach for detection of cluster using POur-NIR method which is an alternative method [PUBLICATION ABSTRACT]
ISSN:1947-5500