Author(s): |
Xianggong Hong, Information Engineering Academy of Nanchang University, Nanchang, Jiangxi Province, 330000 Ping Wang, Information Engineering Academy of Nanchang University, Nanchang, Jiangxi Province, 330000 An Zhao, Information Engineering Academy of Nanchang University, Nanchang, Jiangxi Province, 330000 |
Abstract: |
The Article presents a kind of multi-rough set sorter model and use it in the field of character identification. First, it offers the procedures and steps of building a multi-rough set sorter model: many reduction sets come of the same training sample set through attribute reduction, every reduction set brings forth a regulation assemble, and every regulation assemble corresponds with a sorter. Then the article introduces the methods of Decision fusion of three kind of multi-rough set sorters, then compares and analyzes them by experiments, finally, choose one of them in practical objects. Practice proves this method receives a good identification result.
|