PSO Based Capital Accounting Information Disclosure Model Using Machine Learning Approach for Smart Index System

The efficiency of capital markets and the allocation of social resources depend on the disclosure of capital accounting information. In order to analyze the trend of capital accounting information disclosure, this paper proposed a capital accounting information disclosure model based on immune particle swarm algorithm. There are many factors that influence the disclosure trends of capital accounting information. Prioritizing the level of corporate governance and the level of financial conditions is necessary to develop an impact indicator system for disclosing capital accounting information. A capital accounting disclosure model was constructed to establish the functional relationship between each factor variable and disclosure trends. The particle concentration was maintained through an immune memory and self-regulation mechanism to ensure the heterogeneity of the population, which avoids the traditional drawbacks of the particle swarm optimization algorithm. Finally, parameter estimation of the capital accounting disclosure model is completed. The results show that there are four factors that influence the disclosure trend of capital accounting information, including ownership structure, leverage, growth and audit opinion. The accuracy of the model used in this paper is up to 75%.