A PSO Cluster Technique Based Appraising Model Of The Listed Company Financial Capacities

Date Mining is a new technology coming forth with the development of the artificial intelligence technology and the database technology. It is a complex process of retrieving unknown, valuable models or potentially useful information from mass data, which is widely used in finance, insurance, customer relationship management, manufacturing, retail business and other industries. Clustering is a method commonly used in date mining. The purpose is to divide the data into sub-categories with a certain sense. The similarities between the objects in the same subclasses and the differences between the sub-categories are maximized. Clustering analysis provide a very effective way to mining subgroups which have model in common, and it has been applied widely to marketing management, financial data statistics, satellite imagery, information retrieval and other fields. Clustering Algorithm has division method, hierarchy method, density-based method, grid-based method, model-based method, and fuzzy clustering method (FCM).Particle Swarm optimization (PSO) is one of the main algorithm for swarm group. Swarm is a smart property shown by agents without intelligence when cooperating with each other. PSO is an iterative optimization tool like the genetic algorithm (GA). However, there is no cross-mutation in the process of operation; instead, it's the search for the solution space using the particles'follow of the optimal ones in the space. The process is simple and easy to implement, and the parameters used are concise without complex adjustments. Therefore, the PSO is quickly applied to function optimization, neural networks, fuzzy control systems, data clustering and other fields.This article constructed a critical control clustering model based on particle swarm on the basis of others'work about particle swarm clustering algorithm home and abroad. Each room for the birds to find food can be seen as a feasible solution. There are K cluster centers in each feasible solution; K is the number of categories waiting for clustering. Particles adjust their speeds and positions through recording their best positions and continuously following the optimal location of others. It can finally find the optimal cluster center by means of gradually iteration. Here is the critical control, if the distance between the two particles is smaller than the smallest distance or the preset threshold, the positions of the particles will be redistributed. In this way, it will effectively avoid the emergence of empty cluster and enhance the diversity of particles, effectively prevent the overall search to partial optimum.Particle swarm model uses the dispersion method to compare the particles with the separation and aggregation method to measure the clusters'qualities. The distances of each point to its cluster center approximate normal measurement. Cluster distance within the range of standard deviation measurement and the countdown gathering, the greater the numerical category aggregation is better. We used the clustering interval estimation method of calculation to measure the radius of the cluster center and then used the ratio between the cluster centers and their radiuses to measure the separation.This paper first used the visualization method to verify effectiveness of the the model. 100 Two-dimensional group data points chosen randomly were displayed in the framework of the vertical coordinate plane. There are no distinct categories boundaries while a lot of partial optimal points. Particles found an optimal cluster center through the dynamic display of the search path. Then we used a section of 255 defined three-dimensional data to test the model's search efficiency, and the particles could find the known cluster center within a very short time.It is very necessary To have better financial ability for enterprises to finance in the capital market, because only those ones can receive the trust and support from investors and finally raise fund from the market successfully. Therefore, the scientific evaluation of the listed company's financial position is of great practical significance. Financial ability contains financial solvency capacity, growth capacity, operation capacity, profitability and so on. This paper will first use the swarm cluster model to analyze the four financial capacities of the listed company. Through the construction of the appraising model to the listed companies'financial capacities, we expanded the application of PSO, and also injected new vitality to the traditional financial analysis and provided new ideas for it.The empirical part of our article contains two main aspects. One is the construction of the appraising model to the listed companies'financial capacities; the other is to this appraisal model valid confirmation and the contrast analysis.PSO– based clustering results show that the model has a stronger recognition of the ST enterprises while using the selected 50 financial ratios as input variables. In many of our experiments, the ST alone clustered as one group and there are significant differences in the numbers of ST in the other clusters. "ST"is a special treatment implied to the listed company in the china's open capital market which was abnormal. ST enterprises are those with serious problems in financial situation. According with the core idea of the cluster, we could use the proportion of ST enterprises in the clusters as an important criterion to identify their financial situation.In the traditional evaluations of financial abilities, the weight of each capacity is preset by experience. For example, some scholars believe the ratio of profitability, solvency and growth is 5:3:2.This paper constructed a PSO cluster technique Based Appraising Model of the Listed Company Financial Capacities. We use PSO to cluster the listed company's insolvency, growth, operation, profitability each to identify clusters with different capacities. We determined the ratio of the four financial capacities by using the separation and aggregation degrees between the cluster categories and their contribution to clustering results. Therefore, in the computation of the overall scores, weight has strong scientific value basis and its value could change as well.This paper built a PSO based financial capacity evaluation system using the real data of 52 listed auto manufacturing enterprises on the basis of the PSO cluster technique based Appraising Model of the Listed Company Financial Capacities. The advantages of this paper are the using of real data and scientific computational method and the forefront bionic model and the less using of subjective factors in the entire calculation process. We tested our model's rationality and constructed a financial evaluation model to promote the application of PSO and injected new vitality to the listed company's financial capacity analysis.