论文部分内容阅读
为分析中国股票市场统计特征,利用通过滑动时间窗口建立相关性网络序列,通过超度量矩阵、最小生成树和阈值法等方法转化相关性网络,建立相关性网络的统计特征序列,并分析各个统计特征间的关联与影响。研究结果表明,上证指数收益率对股票相关性网络的聚类系数有负效应,对平均最短路径有正效应,表明中国资本市场上升的动量具有分散化的特征;同时,相关性网络的聚类系数对网络同步性有正效应,平均最短路径则对网络同步性有负效应。
In order to analyze the statistical characteristics of China’s stock market, a correlation network sequence is established by sliding time window, and the correlation network is transformed by hyper-metric matrix, minimum spanning tree and threshold method, and the statistical feature sequence of the correlation network is established. Relevance and influence of features. The results show that the SSE returns have a negative effect on the clustering coefficient of stock-related networks and a positive effect on the average shortest path, indicating that the momentum of China’s capital market is decentralized. At the same time, the clustering of correlation networks Coefficient has a positive effect on network synchronization and average shortest path has negative effect on network synchronization.