An application of logistic regression in cerebral infraction disease detection based on association rules with pre-rough classifier

Environment, customs and health status in northwest minority areas have been studied. We found the critical factors to prevent cerebral infraction. First rough sets theory had been used to reduce the attributes, secondly association rules had been used, finally logistic regression model had been used. The model solved the shortcomings of too many rules that caused by attribute redundancy and reliability framework. The results show that the history of other cerebrovascular disease, alcohol consumption and seasonal change are the significant factors of cerebral infraction.