Abstract

From the cross perspective of media emotion and multilayer network correlation, this paper uses big data technology to mine the data of news reports about bank liquidity, constructs a bank liquidity multilayer network including different media emotional layers, and then analyzes the correlation characteristics of interbank liquidity risk. The main conclusions are as follows: there are banks with higher strength and activity in the bank liquidity multilayer network. These systemically important banks are large state-owned banks and joint-stock commercial banks with large assets. However, among systemically important banks based on strength, large state-owned banks are dominant, and when based on activity, joint-stock commercial banks have more significant influence. The interlayer correlation of the bank liquidity multilayer network is high, and the importance of banks is relatively consistent in different layers. A bank liquidity multilayer network has the characteristics of a small world network with smaller average path length and larger clustering coefficient. The interlayer similarity of the bank liquidity multilayer network is low, and the substitution is poor between positive media emotional network layer and negative media emotional network layer.

Graphic abstract

中文翻译：

---

基于媒体情感的银行流动性多层网络

摘要

从媒体情感和多层网络关联的交叉角度出发，利用大数据技术挖掘有关银行流动性的新闻报道数据，构建包含不同媒体情感层的银行流动性多层网络，然后分析银行间流动性的相关性特征。风险。主要结论如下：银行流动性多层网络中存在强度和活性较高的银行。这些具有系统重要性的银行是大型国有银行和资产较大的股份制商业银行。但是，在基于实力的具有系统重要性的银行中，大型国有银行占主导地位，而基于活动，股份制商业银行的影响更大。银行流动性多层网络的层间相关性高，银行的重要性在不同层次上相对一致。银行流动性多层网络具有世界范围较小，平均路径长度较小，聚类系数较大的特征。银行流动性多层网络的层间相似度低，正向媒体情感网络层与负向媒体情感网络层之间的替代性差。

图形摘要