基于1D-CNN深度学习对科创板信息披露质量的评价研究

Quality Evaluation of Information Disclosure on the Science and Technology Innovation Board Based on 1D-CNN Deep Learning

针对科创板上市公司的信息披露质量评价问题,提出了1D-CNN深度学习模型。使用2022年主板上市公司数据与其对应的信息披露质量评价结果来训练1D-CNN深度学习模型,得到了2022年科创板信息披露质量评价结果。研究结果表明,科创板上市公司信息披露质量在不同行业之间有所差距,其中信息技术行业的信息披露质量水平最好,而医疗保健行业的信息披露质量水平最差;科创板整体的信息披露质量水平是不及主板水平的,但信息技术行业的信息披露质量水平却优于主板水平。造成这一结果的主要原因可能是由于科创板是新兴板块,信息披露体系仍在逐步完善,但由于科创板对科创属性的要求更高,导致信息技术行业的整体信息披露质量水平较高。

Aiming at the evaluation of information disclosure quality of listed companies on the Science and Technology Innovation Board,a 1D-CNN deep learning model was proposed.The 1D-CNN deep learning model was trained using the data of listed companies on the Main Board in 2022 and the corresponding information disclosure quality evaluation results to obtain the information disclosure quality evaluation results of the 2022 STAR Board.The results show that the information disclosure quality of listed companies on the STAR Board varies between different industries,among which the information technology industry has the best information disclosure quality and the healthcare industry has the worst information disclosure quality.The overall level of information disclosure quality of the STAR Board is not as good as that of the main board,but the level of information disclosure quality in the information technology industry is better than that of the main board.The main reason for this result may be that the STAR Board is an emerging sector and the information disclosure system is still gradually improving,but due to the higher requirements of the STAR Board for the attributes of science and innovation,the overall information disclosure quality of the information technology industry is high.