Efficient contrast subspace mining method for categorical data

Florence Sia Fui Sze and Rayner Alfred (2023) Efficient contrast subspace mining method for categorical data.

Text FULLTEXT.pdf Restricted to Registered users only Download (498kB) | Request a copy

Abstract

Mining contrast subspace has recently received attention to identify contrast subspace where a query object is most likely similar to a target class but least likely similar to other class. It has many important applications in various domain such as healthcare, security, finance, and business. Tree-based contrast subspace mining method (TB-CS Miner) has been introduced that is capable to identify contrast subspace of query object in categorical data set. However, the effectiveness of the method has not been evaluated thoroughly. Bedsides, the efficiency of the method in finding contrast subspace has not yet been examined. Real world data sets more often than not containing large amount of data. It is important to have a method that can identify contrast subspace of query object not only effectively but also efficiently on large data set. This paper uses various classification methods to further evaluate the effectiveness of the TB-CS Miner and assesses the execution speed of the method in mining contrast subspace. TB-CS Miner is experimentally shown to be significantly faster and as effective in identifying contrast subspace of query object on real world categorical data sets.

Item Type:	Proceedings
Keyword:	Mining contrast subspace, Contrast subspace, Categorical data, Data mining, Feature Selection, Classification
Subjects:	Q Science > QA Mathematics > QA1-939 Mathematics > QA273-280 Probabilities. Mathematical statistics
Department:	FACULTY > Faculty of Computing and Informatics
Depositing User:	JUNAINE JASNI -
Date Deposited:	11 Aug 2025 16:27
Last Modified:	11 Aug 2025 16:27
URI:	https://eprints.ums.edu.my/id/eprint/44889

Actions (login required)

View Item