Research On Mixed Data Knowledge Acquisition Method Based On Neighborhood Multi-granularity Rough Sets

As an effective data mining analysis tool,rough set theory has been widely used in knowledge discovery,machine learning,image processing and other fields.Attribute reduction and rule acquisition are the core research contents of rough set theory,and many relevant research results have been obtained in the past 30 years.However,with the rapid development of information technology such as communications,sensing and artificial intelligence etc,the characteristic of data attribute in many industries is increasingly showing mixed.Moreover,the amount of accumulated data continues to grow at an alarming rate.Although there is the immeasurable value contained in big data,only by mining the meaningful knowledge from data can we really play the potential value of these massive data.Therefore,how to effectively process these massive mixed attribute data from multi-granularity and multi-level perspectives has become a challenging research topic.This paper mainly studies the knowledge acquisition methods for mixed data based on neighborhood multi-granularity rough set.The main research contents include:Fast attribute reduct algorithm for neighborhood multi-granulation rough sets based on double granulate criterionTo effectively reduce the iterations in computing attribute reduction and realize the quick attribute reduction algorithm,the effect on positive region,caused by different attribute subsets and different neighborhood radiuses,is deeply analyzed based on double granulate criterion.Considering the monotonicity of positive region with the joint function of attribute subset and neighborhood radius,a quick attribute reduction algorithm of neighborhood multi-granulation rough set model based on double granulate criterion is developed.The theoretical analysis and comparable experiments,verify the effectiveness and superiority of the proposed algorithm.Parallel attribute reduction algorithm for neighborhood multi-granulation rough sets using MapReduceIn order to process distributed data with varied data types,and reduce time complexity of attribute reduction algorithm,based on neighborhood multi-granulation rough sets,a parallel reduction algorithm using MapReduce is developed by extracting parallel points from three aspects: hash algorithm,positive region calculation and boundary samples deletion.The comparable experiments under multiple datasets verify the effectiveness of designed algorithm.Rule acquisition for pessimistic neighborhood multi-granularity rough sets based on maximal granuleIn order to further process the numerical or mixed data from the perspective of multi-granularity and multi-level,the pessimistic neighborhood multi-granularity rough sets model is adopted.Through calculating neighborhood multi-granularity condition and decision granular,the redundancy relation of condition granular in the process of rule acquisition is analyzed,on basis of which the redundant condition granular is further pruned.And then,a novel rule acquisition algorithm for neighborhood multi-granularity rough sets based on maximal granule is developed.for the shortcomings of given neighborhood radius in practical applications,an adaptive radius is introduced based on neighborhood multi-granularity rough set model,and the maximum granule based on double granulate criterion neighborhood multi-granularity rough set model is calculated.Further more,a novel rule acquisition algorithm for double granulate criterion neighborhood multi-granularity rough sets based on maximal granule is developed.Comparable experiments also have been realized.