上证综合指数波动情况研究——基于滚动窗口的马尔科夫链预测模型

文章运用基于滚动窗口的马尔科夫链预测模型,对上证综指的变动进行研究,创新的给出概率转移矩阵、极限概率以及预测准确率的时变特征,并给出马尔科夫链预测模型的最优窗口长度和状态定义阀值。研究揭示,大盘波动幅度与大盘的极限概率有着密切的关系;股指期货推出后大盘平盘概率占据主导地位,平稳性显著提高,马尔科夫链预测模型的预测准确率也有了较大提高。

Research on the volatility of Shanghai composite index with the Markov prediction model based on rolling window

Rolling window based Markov chain prediction model was adopted to predict the volatility of Shanghai composite index. The time-varying characteristics of the transfer matrix,limiting probability and forecasting accuracy rate of the Markov chain model were analysed. The optimal window length and the state definition threshold of the Markov chains prediction model were presented. The results indicate that the market volatility and market limiting probability are closely related,that after the introduction of stock index futures the market flat probability occupies the dominant position with increased market stability and prediction accuracy of the Markov prediction chain model.