Improved Density Peak Clustering Algorithm to Optimize Financial Decision Support System

To improve the performance of financial decision support systems. This article proposes a density peak clustering algorithm for deep analysis of financial data. Establish a density peak clustering financial analysis model by extracting features from financial data samples. By calculating the distance between each node and each cluster center point, the sample category is determined based on the set distance threshold. In order to prevent unreasonable distance threshold setting from causing significant disturbance to clustering results, a fruit fly optimization algorithm is introduced to optimize and improve the distance threshold parameters. By continuously updating and optimizing the taste concentration of the fruit fly population, the optimal fitness individual is obtained. The density peak clustering algorithm optimized by the fruit fly optimization algorithm is used to complete clustering analysis and obtain financial data clustering results. The experimental results show that by reasonably setting parameters such as the range of odor change rate in the fruit fly optimization algorithm, the clustering accuracy of all 6 industry samples exceeded 0.9. Compared to commonly used financial data analysis clustering algorithms, the density peak clustering algorithm improved by the fruit fly optimization algorithm can achieve higher clustering accuracy and stability.