Improvement Of Swarm Collaborative Intelligent Optimization Algorithm And Its Application Research

The main contents and inno Optimization problems exist extensively in practical engineering problems and scientific research.The optimization problem has the characteristics of large scale and high dimensionality of the solution space.The traditional optimization algorithm solves these problems which has problems such as high computational complexity and long time.Swarm intelligence algorithm has become a new research direction for solving optimization problems because of its advantages of fewer parameters,simple model and easy implementation.With the rapid development of the artificial intelligence,e-commerce and mobile internet finance continue to generate data at all times.Data mining technology has received increasing attention in many fields.Clustering technology is an important branch of data mining.Under unsupervised conditions,it is used to mine potential data structures,which has become a research hotspot in the field of artificial intelligence.The clustering by fast search and find of density peaks algorithm is a new and very competitive in the clustering algorithm,which has been widely recognized in various fields.However,it still has the drawback of manually setting parameters.This paper not only takes the cuckoo search algorithm as the main research object,researches and improves it,but also improves the shortcomings of the density peak fast search clustering algorithm.vations of this article are as follows:(1)To solve the problem that the cuckoo search algorithm converges slowly,when dealing with complex functions,and it has low searching accuracy and poor algorithm stability when dealing with multidimensional data,a novel dual evaluation strategy-based cuckoo search algorithm with dynamic self-adaptive step size is proposed.The algorithm introduces a dynamic adaptive step length mechanism and a double evaluation strategy.In the dynamic step size,the learning factor acceleration algorithm searches the speed in the solution space.In the early iterations of the algorithm,the column-by-column sorting strategy in the dual evaluation strategy is quickly located in the global search and it introduces the dynamic discovery probability to increase global search capability.(2)To solve the problem that clustering by fast search and find of density peaks algorithm has the disadvantages of manually setting the cutoff distance dc and the Euclidean distance can not accurately reflect the similarity between data,an improved cuckoo search optimization-based density peak clustering algorithm is proposed.The improved algorithm chooses the cuckoo search algorithm to optimize the truncation distance and introduces the cosine similarity.It combines the direction with the actual distance to better distinguish the attribution of data points in the two types of intermediate regions.Simulation results show that the improved density peak fast search clustering algorithm has better clustering performance.(3)It clusters the personal credit data of the bank based on the improved cuckoo search optimization-based density peak clustering algorithm.The simulation experiment results show that the method proposed in this paper can effectively analyze and predict the bank's personal credit default situation and help the bank's credit department make a reasonable decision.