The Research Of Clustering Analysis Based On Coupled DNA-GA-P Systems

According to the working principle of organelles and cell membranes in organism,P system can operate in a maximum parallel mode,and its computing ability is equivalent to Turing machine.At present,it has been used to deal with data mining problems by scholars.DNA genetic algorithm simulates the genetic information expression mechanism of biology.This kind of information expression process also occurs in eukaryotic cells.The effective of coupled DNA-GA-P system can not only preserve the distributed parallel computing ability of P system,but also integrate the rich object expression mechanism and gene level operation of DNA genetic algorithm,and expand the object and rule of P system computing model.The rule expression provides a new dynamic evolution model for the existing P system,and further expands the problems that P system can handle.Nowdays,with the explosive growth of mobile data,the traditional data processing methods can not meet the needs of data processing.Data mining is born from this.Clustering analysis is an important research method in the field of data mining.But the current clustering method has some shortcomings.In addition to improve the algorithm itself,we can also use other optimization methods to further optimize the clustering process.So we need to combine new computing methods or models to further optimize the clustering effect.The research of new methods and models is an important topic in the field of data mining.The main research contents of this paper are as follows:(1)Construct the coupled DNA-GA-P systemBased on the biological knowledge of P system and DNA genetic information,a new coupled DNA-GA-P system is constructed.At the same time,based on the coupled DNA-GA-P system,four extended coupled DNA-GA-P systems are proposed,which are the coupled DNA-GA-P system with directional communication and probabilistic evolution rules,the coupled DNA-GA-P system with membrane division / membrane dissolution rules,the coupled DNA-GA-P system based on chain topology and the coupled DNA-GA-P system based on self-assembly concept.The calculation ability and efficiency of the proposed P system are proved in the paper.(2)Four kind of clustering algorithms are improved,and four coupled DNA-GA-Psystems are used to implement the improved clustering algorithm,respectively.A.Fuzzy c-means clustering algorithm based on coupled DNA-GA cell-like P system.The distance calculation method which based on weight mean concept is used to calculate the objective function of fuzzy c-means clustering algorithm.The new coupled DNA-GA-P system is used to realize the clustering process.The improved algorithm is further optimized by the global search ability and the ability to jump out of the local optimum of the coupled DNA-GA-P system.The UCI datasets are used to verify the performance of the improved algorithm.B.Density peak clustering algorithm based on coupled DNA-GA tissue-like P system.In this method,the K-Nearest Neighbor and Shannon entropy are simultaneously used to calculate the density matrix of data points.The coupled DNA-GA tissue-like P system is used to implement the clustering process.The new tissue-like P system can not only improve the efficiency of the algorithm,but also reduce the complexity of the algorithm.Finally,experiments are carried out on artificial data set and real data set,respectively.C.Consensus fuzzy K-modes algorithm based on coupled DNA-GA chained P systemAll attributes of fuzzy K-Modes algorithm are solved in a balanced way based on intuitionistic fuzzy set(IFS)and kernel trick.And the noise robustness of the algorithm is improved.Then,the consensus fuzzy K-modes algorithm which combined the advantages of the improved fuzzy K-modes algorithm and the other two kind of Kmodes algorithm are proposed.Coupled DNA-GA chained P system is used to implement the proposed consensus fuzzy K-modes algorithm to prevent the clustering algorithm falling into local optimum,and realize the implicit parallel clustering process.D.Multi-view spectral clustering algorithm based on coupled DNA-GA population P system.A new automatic weighted multi-view consensus clustering algorithm based on KNN and graph is proposed.On the one hand,K-nearest neighbor is used in the initialization of data representation matrix(similarity matrix).On the other hand,the similarity matrix is used to instead of the original data object to learn the consensus matrix.The similarity matrix will be updated in the iteration process.Then,in the process of generating consensus matrix,the system automatically generates weight for each view,and updates the weight information of each view synchronously in the later update process.Finally,when the consensus graph converges,the spectral clustering algorithm is performed on the consensus graph,and the final multi view clustering result is obtained.The multi view map clustering process is completed in the coupled DNA-GA population P system according to the specific rules.The maximum parallelism of the system can further improve the efficiency of the algorithm.(3)Fuzzy c-means clustering algorithm based on coupled DNA-GA-P cell-like P system and multi-view spectral clustering algorithm based on coupled DNA-GA population P system are applied to image segmentation and text clustering,respectively.In summary,this paper proposes a new coupled DNA-GA-P system,and proposes four extended coupled DNA-GA-P systems which based on the system definition.At the same time,four extended systems are used for solving four improved clustering algorithms,respectively.Finally,two clustering algorithms based on the coupled DNAGA-P system are used in the practical application of image segmentation and text clustering.