Research And Application Of Partitioning Clustering Algorithm Based On Tissue-like Membrane System

Membrane computing provides us with a new computing model for mass data processing.Membrane computing has the characteristics of distributed,maximum parallelism,mass storage and uncertainty.The application of membrane computing in clustering will greatly improve the clustering efficiency on the premise of guaranteeing the quality of clustering.Partition clustering algorithm has high time efficiency because of its close linearity of time complexity.At the same time,it has been widely used because of its simple idea and its ability to deal with large-scale data objects effectively.However,the clustering results of partition-based clustering algorithms are often affected by the initial values,which need to be given in advance and highly dependent on the initial clustering centers.At the same time,the algorithm is sensitive to noise.Individual abnormal data may seriously affect the final result of clustering,and may increase the number of iterations,thus affecting the convergence speed of clustering algorithm.Generally,the algorithm is easy to fall into local minimum.Firstly,the research background and significance are introduced in detail.The development status of membrane computing and partitioning clustering at home and abroad is reviewed.The research status and development of the combination of membrane computing and partitioning clustering are introduced.Then,the tissue-based membrane system and partitioning clustering are summarized.In this paper,the theory and application are studied.Firstly,a tissue-based membrane system based on annular membrane structure is proposed,and the clustering quality of the system is studied by a simple example.Secondly,the partition clustering algorithm is improved and applied in combination with the tissue-based membrane system based on annular membrane structure and network intrusion detection.The main research contents are as follows:Firstly,the tissue membrane system was improved by introducing annular membrane structure into the tissue membrane system,and a tissue membrane system based on annular membrane structure was designed.In this membrane system,each membrane has a single role,a clear division of labor,and the execution efficiency is greatly guaranteed.Secondly,an IKM algorithm based on CSTP system is proposed,which improves the traditional method of random selection of initial clustering centers.For any data set,the K-means algorithm can determine the initial clustering centers at a very fast speed,and a mutation strategy is proposed to optimize the operation results of the algorithm.Thirdly,HCBP algorithm based on CSTP system is proposed,which improves the selection of clustering centers in the iteration process of K-center algorithm,updates the calculation formula of cost function,and combines K-center algorithm with K-mean algorithm to further optimize the results of K-mean algorithm.And a hybrid partition clustering algorithm based on the annular structure of tissue membrane system is proposed.Fourthly,CSTP-HCBP algorithm and CSTP-IKM algorithm are applied to network intrusion detection to detect the experimental data KDDCUP99.Intrusion network detection is to collect important data information from websites for technical analysis,so as to determine whether it is normal or intrusion attack.