Classification rule extraction based on the rough concept lattice

Abstract

Purpose

The purpose of this paper is to examine the important application value of extending the concept of classification rule, so that it can describe and measure the uncertainty of classification knowledge.

Design/methodology/approach

The rough concept lattice (RCL), which is an effective tool for uncertain data analysis and knowledge discovery, reflects a kind of unification of concept intent and upper/lower approximation extent, as well as the certain and uncertain relations between objects and attributes.

Findings

A classification rules extraction algorithm, extraction algorithm of classification rule (EACR), based on the RCL is presented by adapting the rough degree to measure uncertainty of classification rule. The algorithm EACR is experimentally validated by taking the star spectrum data as the decision context.

Practical implications

An efficient way for classification rule extraction is provided.

Originality/value

The algorithm EACR based on the RCL is presented by adapting the rough degree to measure uncertainty of classification rule.

Citation

Hai‐feng, Y., Ji‐fu, Z. and Li‐hua, H. (2010), "Classification rule extraction based on the rough concept lattice", Kybernetes, Vol. 39 No. 8, pp. 1336-1343. https://doi.org/10.1108/03684921011063637