【摘要】 运用分位数半径的动态K-means状态划分，建立了一个改进的期望贝叶斯(Expected Bayes, E-Bayes)估计模型。通过提取有效的时序信息来构造新的观察值向量，并对E-Bayes估计进行权重调整，进一步提高了模型的预测精度。实证分析结果表明，在上证指数的区间预测中，相较于单一的E-Bayes方法，改进模型的预测准确率提升了29.16%。更多还原

【Abstract】 Using the dynamic K-means state partition of quantile radius, an improved Expected Bayes(E-Bayes) model is established. By extracting effective time series information to construct a new observation vector, and adjusting the weight of E-Bayes estimation, the prediction accuracy of the model is further improved. The empirical analysis shows that, compared with the single E-Bayes method, the prediction accuracy of the improved model has been increased by 29.16% in the interval prediction of the Shanghai Stock Exchange Index.更多还原