Title: Research on financial irregularities identification: a machine learning perspective

Authors: Kuang-Cheng Chai; Jiawei Zhu; Yang Yang; Hao-Ran Lan; Yang-Lu Ou; Qiang Li

Addresses: Business School, Guilin University of Electronic Technology, Guilin, Guangxi Province, China ' Business School, Guilin University of Electronic Technology, Guilin, Guangxi Province, China ' Business School, Guilin University of Electronic Technology, Guilin, Guangxi Province, China ' Business School, Guilin University of Electronic Technology, Guilin, Guangxi Province, China ' Business School, Guilin University of Electronic Technology, Guilin, Guangxi Province, China ' Business School, Guilin University of Electronic Technology, Guilin, Guangxi Province, China

Abstract: In the era of big data, data-driven analytics can generate many meaningful insights. With the gradual maturity of artificial intelligence (AI) algorithm, it can help solve problems that are difficult to identify in the financial field. Through theoretical analysis, this paper constructs feature engineering of multiple internal and external factors that affect corporate financial irregularities, and then automatically identifies Chinese listed companies with financial irregularities based on machine learning algorithm. In this paper, we verify the effectiveness of SMOTE algorithm in improving the imbalance data of financial irregularities of Chinese listed companies and use LightGBM algorithm to sort the ten factors of characteristic importance of financial irregularities of Chinese listed companies. This paper provides a new way to detect financial irregularities for financial regulatory authorities and a paradigm for the applying of AI in the financial field.

Keywords: financial irregularities identification; machine learning; SMOTE algorithm; China.

DOI: 10.1504/IJIMS.2022.128641

International Journal of Internet Manufacturing and Services, 2022 Vol.8 No.4, pp.383 - 399

Received: 09 Nov 2021
Accepted: 05 Aug 2022
Published online: 31 Jan 2023 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article