Research On Maximal Association Rule Mining And Applications Based On Temporal Parameter-Taxonomic Soft Sets

Data mining generally refers to all analytical methods and algorithmic procedures related to knowledge discovery in databases,which aims to automatically identify and extract various forms of knowledge such as association rules,potential principles or patterns hidden in the collected data.The obtained knowledge can be used to facilitate decision-making,control and behavior in various practical activities.In the era of big data,how to extract valuable knowledge from massive data has attracted unprecedented attention,due to the challenge of data explosion and knowledge shortage.With the emerging of data mining,interdisciplinary research involving mathematics,computer science and artificial intelligence has been greatly promoted.Association rules mining is one of the most critical technologies in data mining.Its purpose is to find frequent item sets and potential association rules between different item sets from transaction data sets.Maximal association rules,as a useful supplement to regular association rules,can help to reveal the specific inner associations that are ignored in the regular association rules mining.Maximal association rules mining can effectively reduce the number of redundant rules and eliminate invalid rules.It should be noted that soft set theory provides a powerful mathematical tool for coping with uncertain information since it can describe uncertainty by considering both the universe of discourse and the related parameter space.This theory enables more comprehensive mathematical description and information processing in the process of data mining.In addition,traditional association rules mining technology often ignores the influence of temporal factors.Nevertheless,in many pragmatic scenarios,the data will change dynamically over time,and the intrinsic associations carried by the data will vary accordingly.This paper defines and studies the temporal soft sets,temporal parameter-taxonomic soft sets and their basic properties.We propose a novel rule pattern named temporal maximal association rules.We also propose the soft sets description method of temporal maximal association rules and maximal association rules mining algorithm based on temporal parameter-taxonomic soft sets.In addition,we study the operation efficiency,threshold sensitivity and redundant processing capacity of the algorithm.We also apply the proposed mining method to explore the global climate change problem.The core contents of the paper are as follows:(1)By employing the Web of Science database literature analysis tool and VOSviewer literature visualization software tool,we systematacially study the knowledge structure and research progress of journal articles,which are included in Web of Science core collection database on data mining based on soft sets.We also use literature co-occurrence clustering analysis function to analyse these journal articles from different angles,such as keywords,researchers and countries.(2)Soft sets,soft subsets,soft equality relationship and the basic operation of soft sets are briefly introduced.We also discuss the relationship between soft sets and information systems.In addition,this paper systematically studies the basic knowledge of regular association rules and maximal association rules,and summarizes the related theories of temporal association rules mining.Finally,the existing methods classification of temporal association rules mining is given.(3)Combining parameterization and granularity methods,the concepts of temporal soft sets and temporal parameter-taxonomic soft sets are proposed.A novel rule pattern named temporal maximal association rules is introduced,and soft sets description method of temporal maximal association rules is developed.In addition,maximal association rules mining algorithm based on temporal parameter-taxonomic soft sets is constructed.The effectiveness of the algorithm is verified via numerical examples and comparative analysis with several existing mining algorithms.(4)An application of the proposed temporal maximal association rules mining method to global climate change research is explored.By data mining and knowledge discovery conducted on the data sets regarding global extreme weather and climate events from the international emergency and disaster database,the correlations and potential principles of extreme weather and climate appearing worldwide are revealed.