Construction of Control Rights Allocation Index of Listed Companies Based on Neural Network and Machine Learning

Abstract

Control power is a core issue that every listed company pays great attention to. The company’s shareholding structure directly affects the allocation of control rights. Therefore, the shareholding structure of listed companies is analyzed, and various factors related to the allocation of company control rights are discussed. It is very important to build indicators of control allocation of listed companies and improve the governance model of listed companies. Based on this, this article proposes to use neural networks and machine learning techniques to build related models and solve related problems. This article takes the control allocation index of listed companies on the SSE and SZSE platforms under good securities’ market conditions as the research object and takes the stock holding allocation of listed companies as a reference for the control allocation index. Combine sliding removal technology and approximate entropy with sample entropy, select the sliding window and sliding step size as 21 data, keep the sliding window unchanged, and calculate the approximate entropy and sample entropy of the sequence after removing 21 data for each sliding value to analyze the correlation between the rate of return, complexity, and effectiveness. The results of the study show that the mean and median of the majority shareholder’s equity pledge behavior are 0.249 and 0, respectively, and the mean and median of the majority shareholder’s equity pledge ratio are 0.147 and 0, respectively, indicating that 24.9% of the companies in the sample have major shareholder equity. Pledge is limited by sample data, and the proportion of major shareholders’ equity pledge is moderate, which means that there is a certain gap in the quality of internal control between companies.