Abstract
- II -
Abstract
In recent years, with the development of informationization, the enterprise
information management has been drawing growing attention. The complicated and
changeable social economic environment, however, increase the difficulty of enterprise
risk management. It is an objective how to establish an efficient risk management system
for an enterprise finance early-warning. The characteristic and complexity of the financial
early warning problem, makes us very difficult to process massive data with the traditional
statistics analytical method. In present, how to screen useful information from the business
activities of enterprises is an urgent problem to solve for the financial risk early-warning
management. Data mining technology improvement and innovation including good
performance analysis of massive data research is good for solving above issue. In this
paper, the data mining technology is applied to research and analysis of the enterprise
financial risk for early-warning, which is very important on the theoretical and practical
significance. The main contents of this paper are as follows:
Firstly, this paper focuses on the analysis and early-warning of financial risk,
including the reasons and characteristics of the financial risk, and financial risk for early
warning method.
Secondly, a more efficient data mining algorithm based on a frequent item sets and a
parallel packet P-Apriori is proposed. In this algorithm, (K-1)-frequent item sets is divided
into groups by certain rules, each group of (K-1)-frequent item sets generates K-frequent
itemsets directly and then combine them. So this will reduce a lot of judgement attempt
when the self-connection and can provide parallel processing capabilities to solve
connection and pruning action, reducing the waiting time and improve the search speed of
frequent item sets. Experiments show that the improved algorithm has greatly improved in
performance. We fulfilled this to provide technical support for the further research on
financial risk early-warning.
Thirdly, the decision tree C4.5 algorithm is analyzed and studied to take the virtual
financial data of listed companies as the training sample data, and show the process of
building financial risk early-warning model.
Fourthly, financial risk analyzing system is built by the P-Apriori algorithm of
association rules for the purpose of the financial index data, processing more convenient,
which provides convenience and reliability for the study of financial risk early warning.
Through the analysis of the 66 ST companies and 499 financial index data mining have
found 11 key indicators 26 frequent financial risk indicators, these indicators can reflect
万方数据
剩余81页未读，继续阅读
资源评论
