WSEAS Transactions on Computer Research
Print ISSN: 1991-8755, E-ISSN: 2415-1521
Volume 12, 2024
Corporate Accounting Management Risks Integrating Improved Association Rules and Data Mining
Author:
Abstract: With the development of the times, enterprises need to face more data in operational decision-making. Traditional data analysis strategies cannot handle the growing amount of data well, and the accuracy of analysis will also decrease when faced with uneven data types. The research uses a corporate accounting management risk analysis technology that combines big data algorithms and improved clustering algorithms. This method combines big data processing ideas with a clustering algorithm that incorporates improved weighting parameters. The results show that on the data sets DS1, DS2, and DS3, the NMI values of the GMM algorithm are all 0; while the NMI values of the MCM algorithm correspond to 0.9291, 0.9088 and 0.8881 respectively. At the same time, the Macro-F1 values of the Verify2 algorithm correspond to 0.9979, 0.9501, and 0.9375 respectively, and the recognition accuracy of the data remains above 85%. In the running time comparison, when the number of samples in the data set reaches 5,000, the calculation time of the Verify2 algorithm remains within 5 seconds. In terms of practical application results, the study selected the profitability risk indicators of 40 companies for analysis. After conducting risk ratings, it can be seen that companies No. 5, 6, 7, and 39 have the highest risk levels, and companies No. 33 and 34 have the highest risk levels. The lowest level. After conducting risk assessments on the 40 selected listed companies, the risk level of net asset income of each company remained at level 5, and the risk level of earnings per share remained at level 3. The above results show that this technology has good performance in terms of calculation accuracy and calculation time, can assess enterprise risks, and can provide data support for enterprise operation decisions.
Search Articles
Keywords: Data mining, Big data, Clustering algorithm, Risk assessment, Association rules, Enterprise operation decisions
Pages: 348-358
DOI: 10.37394/232018.2024.12.34