摘要
研究采用图卷积神经网络对投资者网络互动平台的效能进行诊断与优化,构建了包含信息透明度和互动性的三级指标体系,并与线性降维方法比较。实证研究基于“上证e互动”和“互动易”平台2011年至2023年的交流数据,结果显示,图卷积方法在市场表现解释力上具有优势。基于此,研究建议企业提升信息管理与沟通能力,政府加强信息披露监管,推动数字化转型和技术创新,提升市场透明度,增强投资者信心,促进资本市场健康发展。
This study applies a graph convolutional network(GCN)to diagnose and optimize the service effi ciency of investor online interaction platforms.A three-tier indicator system incorporating information transparency and interactivity is constructed and compared with linear dimensionality reduction methods.The empirical analysis is based on interaction data from the“SSE E-Interaction”and“SZSE Interactive Easy”platforms from 2011 to 2023.The results demonstrate that the GCN method has superior explanatory power in assessing market performance.Based on these findings,the study recommends that companies enhance their information management and communication capabilities,while regulatory authorities should strengthen oversight of information disclosure,promote digital transformation,and foster technological innovation.These measures aim to improve market transparency and stability,bolster investor confi dence,and support the sustainable development of capital markets.
作者
李国文
丁正扬
冯钰瑶
LI Guo-wen;DING Zheng-yang;FENG Yu-yao
出处
《科学决策》
2025年第1期158-177,共20页
Scientific Decision Making
基金
国家重点研发计划(2021YFF0900800)
国家自然科学基金青年项目(72201287,72301271)。
关键词
投资者互动平台
效能诊断
线性降维
图卷积神经网络
investor interaction platform
efficacy diagnosis
linear dimensionality reduction
graph convolutional network
