Optimal Scale Selections And Knowledge Acquisition In Generalized Multi-scale Decision Tables

Granular computing(GrC),which simulates the natural mode of human thinking,is a discipline that specializes in the study of theories,methods,techniques,and tools.GrC is based on granular structure of thinking,problem solving methods,and information processing modes.The emergence of GrC has changed the traditional concept of computing.It processes data information scientifically in a more reasonable and simple way,and shows its unique computing advantages in the era of big data.In the field of data processing and knowledge discovery,GrC has developed into a more focused topic,and its appearance provides researchers with more advanced ideas in data mining.Agriculture,production and management produce a large amount of agricultural data information,which is complex,uncertain,incomplete and redundant.In terms of agricultural informatization,using GrC to extract useful information from agricultural data is a new research hotspot,and it is helpful in promoting the rapid development of agricultural informatization.This dissertation presents the optimal scale combination and knowledge acquisition in generalized multi-scale decision systems.The main contents are organized as follows.Firstly,problem of optimal scale combination selections based on dual probabilistic rough set model in generalized complete multi-scale decision systems is discussed.By using dual probabilistic rough set model,four notions of optimal scale combination selections with variable precision_ ? in inconsistent generalized multi-scale decision systems are defined.Relationship between the above mentioned and the existing concepts of optimal scale combinations is given under certain conditions.Attribute reduction in system is calculated and the IF-THEN rules are carried out.Secondly,problem of optimal scale selections based on dual probabilistic rough set model in incomplete multi-scale decision systems is studied.Definitions of incomplete multi-scale decision systems and optimal scale selections are reviewed.Notions of optimal scale selections under the variable precision ? in inconsistent incomplete multi-scale decision systems are defined.Attribute reduction of decision system is calculated and rules are extracted.Finally,method of knowledge acquisition in inconsistent information systems with multi-scale decisions is discussed.Multi-scale generalized decision is introduced into generalized multi-scale decision systems for the first time.Notions of two optimal scales are defined to keep generalized decision unchanged under the finest scale in inconsistent information systems with multi-scale decisions,and rules are extracted.