Research Article
Research on Recognition of Whitewashing Degree of Financial Statements Combined with Excess-MAD Algorithm and Clustering Algorithm Model
@INPROCEEDINGS{10.4108/eai.26-5-2023.2334343, author={Tingyu Luo}, title={Research on Recognition of Whitewashing Degree of Financial Statements Combined with Excess-MAD Algorithm and Clustering Algorithm Model}, proceedings={Proceedings of the 2nd International Conference on Mathematical Statistics and Economic Analysis, MSEA 2023, May 26--28, 2023, Nanjing, China}, publisher={EAI}, proceedings_a={MSEA}, year={2023}, month={7}, keywords={degree of whitewashing of financial statements; benford's law; k-means model; excess-mad method; gaussian mixture model}, doi={10.4108/eai.26-5-2023.2334343} }- Tingyu Luo
Year: 2023
Research on Recognition of Whitewashing Degree of Financial Statements Combined with Excess-MAD Algorithm and Clustering Algorithm Model
MSEA
EAI
DOI: 10.4108/eai.26-5-2023.2334343
Abstract
Using the improved Excess-MAD algorithm based on Benford's law to establish financial statement whitewashing identification indicators, screen out 15% of the financial statement data available from all listed companies in the Chinese capital market from 2000 to 2021, about 14,000 pieces, and divided into three groups. Then, according to relevant financial theories, 11 financial indicators of 5 theoretical analysis dimensions are selected to establish a preliminary indicator set, and 8 indicators are reserved for modeling after being screened by Kendall correlation analysis. Finally, K-means, Gaussian Mixture Model and other clustering methods were used to establish the final binary clustering model, and Akaike information criterion and Bayesian information criterion were used for extended evaluation. The results of the model run show that the clustering algorithm model established through this process can well identify the degree of financial whitewashing of the company's statements.