Work

Attribute Reduction Based on Sorting and Incremental Method

Year: 2015
Type: article
Abstract: Positive region is one of the core concepts in rough set theory.Time complexity of computing the positive region can directly affect other algorithms.In this paper, a new algorithm for computing equivalence classes based on generalized quick sort and insertion sort is provided and its time complexity for computing U/C is cut down to O(|C||U|) compared with other traditional algorithms.On this basis, an algorithm for fast computing positive region by adding identifier atttibutes to sorted decision tables is proposed and its time complexity for computing POS C (D) is cut down to O(|C||U|) accordingly.By using this foundation, two incremental algorithms for fast computing positive region are designed.They make full use of equivalence classes and positive region which already exist to compute positive region incrementally.Thus, these algorithms can reduce computational work signicantly and get higher efficiency.Among them, the algorithm based on multiway tree is the most efficient.Its time complexity for computing POS C (D) is far less than O(|C||U/C|).Finally, a attribute reduction algorithm based on incremental method is given and its time complexity is O(|C| 2 |U/C|).The example analysis and experimental result show that the attribute reduction algorithm of this paper are feasible and efficient.
Author: Wang Biqing
Institution: Tongling University
Language: English
Open Access status: diamond
OpenAlex ID: W2238327984
FWCI:
Cites: 8
Cited by:
Related to: 10