Research About Brain Science Data Based On Group Search Optimization And Rough Set

fMRI makes it possible for humans to explore the interior of brains without damage. The technology is applied to report the status of the brain neuron and has the potential in terms of location of encephalic region. Data collection is a series of reaction images from different regions faced with specific things during a period, which results in complex regions and huge data. Therefore, reduction on huge amount of data is needed to explore the state of brain activity.To solve the problem of huge data and inaccuratnce which leads to difficulty to analyze of brain data, rough set theory is adopted to solve uncertainty, the principle and reduct of which are researched as well. The main topic is to execute reduct and obtain rules without changing decision ability. To overcome the weakness of single reduct, long running time and difficulty to obtain minimal reduct by rough set, this paper applies group search optimization(GSO) to rough set theory, proposes a rough set based on group search optimization-reduct based on multi-knowledge extraction. GSO is a kind of swarm intelligence algorithm based on animal swam behavious. Similar to seeking food, the algorithm seeks for the optima in the searching space by means of sharing information among members.Particle swarm optimization rough set is compared with the algorithm in this paper on UCI data sets to prove that the algorithm works better. Furthermore, the paper introduces the principle of fMRI and conditions to obtain brain data. The fMRI data in the experiments is obtained from memory tests with outside stimulation which consists of three analyzing pattern. The algorithm is applied to reduct of different brain region data in order to identify activation area in the brain under different kinds of circumstances.