Research And Application Of Frequent Pattern Mining Algorithm Based On Tissue-like P System

Membrane computing is a new branch of natural computing,emphasizes seeking inspiration in the structure and functions of living cells,and its model is often called the P system.A P system mainly consists of three parts: Membrane structure,multiple sets of objects,and evolution rules.It has the equivalent computing power as the Turing machine.Frequent pattern mining is a very important task in data mining,mining patterns that frequently appear together.Associations discovered between the frequent patterns are widely used in various recommendation and prediction models.Nowadays,in the "big data era",massive data is very easy to obtain.The problems of massive data storage and insufficient computing power bring huge challenges to frequent pattern mining.Therefore,it is necessary to apply new computing models and incorporate new improvements.Combination of P system and data mining is not only an extension of the application of P system,but also provides new ideas and methods for the development of data mining.The thesis first introduces the research background and significance,reviews the research status and development trends of membrane computing and frequent pattern mining.The definition of tissue-like P system and frequent pattern mining,the algorithm of frequent pattern mining are introduced in detail,and then the structure and innovation of the thesis are summarized.New membrane computing models are proposed and combined models with frequent patterns mining algorithm,finally applied them to two practical applications.The main contents are as follows:1.Based on tissue-like P system and function of cell division and differentiation,a new division and differentiation adaptive tissue-like P system(DATP)is proposed.The register principle is used to verify the computing power of the system.The adaptive division and differentiation rules can reduce the use of resources.Based on biological enzymes and channel feedback mechanisms,two new P system models,EDATP and FDATP are proposed.The system operation mechanism and evolutionary rules are designed to optimize system structure.2.A direct membrane algorithm ETP-EL based on EDATP is proposed.A pruning strategy is proposed using the containment property and membrane system to reduce the number of judgments and objects used in the system.By using the parallelism and large-scale nature of the P system,the time complexity of the algorithm is greatly reduced,and the effectiveness of the system is verified by experiments.Finally,the importance of threshold selection and the effectiveness of pruning strategy are analyzed by using four datasets.3.A PDT-VTK algorithm based on the FDATP system and VTK is proposed,and the rules and system structure are designed to implement the algorithm's operation.Using difference set to reduce the memory consumption of algorithm in dense data and improve the time efficiency.The format conversion condition is designed to make algorithm automatically selects the appropriate data structure in different situations.UCI,PUMSB and real datasets are used to verify the effectiveness of the improved algorithm.4.Applying ETP-EL algorithm to microblog user recommendation application,designing a microblog user recommendation system based on association rules,and recommending relevant ID to users based on their microblog information.The PDT-VTK algorithm is applied to the supermarket shelf layout application.Frist processing and classifying the original information of the supermarket sales record,and give a reasonable supermarket shelf layout recommendation with the correlation between the product categories.