Research On Particle Swarm Optimization And Rough Sets Reduction

Rough set theory is an effective method of multi-attribute and uncertainty handling com plex systems, this method does not require any prior information and data collection issues ne ed to be addressed outside of the uncertainties described in its issue or treatment is more obje ctive. Knowledge reduction is one of the important researches on rough set theory, its main p urpose is to remove data redundancy, while it makes the classification ability of the original d ecision-making information systems remains unchanged.However, how to get the minimal reduction have been the key issues for researchers. Due to find the minimum attribute reduction has been proved to be NP-hard problem, which will allow random search strategy algorithm(eg: particle swarm optimization, etc.) the introductio n of attribute reduction, which became the improvement of existing reduction algorithm is an effective solution. Due to the principle of particle swarm algorithm is simple, easy to implem ent and has a strong versatility and global optimization of, so it was introduced to attribute re duction algorithm, but the existence of the particle swarm algorithm itself is poor local search capability, easy to fall into the global optimum search accuracy is not high defect, therefore, t he need to improve it accordingly. In addition, based on the traditional attribute importance of attribute reduction process is complex and cumbersome, attribute reduction process can be ca rried out to its improvement, and finding the right method and particle swarm optimization co mbined in order to achieve quick attribute reduction.First, this paper briefly analyzes attribute reduction algorithms, comparing their advantag es and disadvantages. Meanwhile, in every attribute reduction based on important degree of r eduction calculation process, the process of adding the greatest attributes of a current attribute importance to the kernel, and then propose appropriate measures to improve the use of routin g congestion control algorithm Thought reduction process can be simplified by adding contro l mechanisms in the process of reduction of by adding attributes for each single greatest attrib ute importance to nuclear transformation for each and then add more conditions to determine whether the reduction and add fallback mechanism to ensure that the reduction can be acquire d in order to achieve improved speed reduction. Finally, the improved process of attribute red uction algorithm and particle swarm optimization combined together, the proposed attribute r eduction algorithm based on improved particle swarm optimization and rough sets attribute i mportance of global use of this algorithm is mainly particle swarm optimization, simple and e asy to realize the advantages and improved combination of attribute reduction algorithm to ac hieve access to the best attribute reduction purposes.Finally, after many experiments verified, the paper’s "rough set based on PSO and Impro ved Attribute Reduction Algorithm" attribute reduction can improve the speed, faster and more accurate access to attribute reduction.