Research On Meta-Heuristic Clustering And Its Application In Stock Market Analysis

Under the background of big data era,with the ever-increasing value created by data,it has become a more important strategic asset in the financial industry.People expect that data will bring better service to the financial industry.However,how to remove unnecessary information from a large amount of data and select the necessary parts from the data and analyze the data according to the objective needs,these problems have aroused the concern of the majority of scholars.Because stock data has many characteristics,such as complex structure,variety and redundant data,how to represent and analyze data more efficiently becomes a challenging issue.The cluster analysis is the object in accordance with certain rules of cluster or class division,the essence of the study is concerned about the degree of similarity or correlation between objects.If there is a high degree of similarity between the same category,while large differences between categories,can be considered to have achieved relatively good clustering results.According to the different structure of the data,the existing clustering algorithm will produce different clustering results.In the process of securities investment,people pay more attention to how to better integrate the basic factors of the stock market to cluster them,and improve the measurement ability of the similarity of the sample securities.This topic is proposed in this context.In this paper,meta-heuristic algorithm and clustering analysis method are introduced in detail.The basic concepts and definitions of mathematical principles and process steps are analyzed in detail.The corresponding knowledge reserves are provided for the following algorithm improvement.Secondly,in order to solve the problem of convergence speed and accuracy in traditional fruit fly optimization algorithm,a fruit fly optimization algorithm based on combinatorial mutation is proposed.By introducing combinatorial variation strategy,the flight direction of the population is updated and the population escape ability is enhanced.It provide theoretical basis for the improvement of the clustering algorithm in the following text.Then,an improved fruit fly optimization algorithm based on combinatorial mutation is used to improve the traditional AP algorithm.An adaptive AP algorithm based on CMFOA optimization is proposed.Using the Silhouette index to evaluate each validity,which can search parameter space fast,enhance the global exploration ability of the algorithm,strengthen the local optimization ability of the algorithm,obtain the best clustering structure and improve the clustering performance of the algorithm.Finally,the research results of this subject are validated and analyzed through examples,and the improved algorithm is verified by using multiple evaluation indexes such as Fm and Sil simultaneously.The experimental results show that the algorithm improves the performance of clustering and further improves the accuracy of clustering results.