【摘要】 研究了异方差混合双自回归模型(Heteroscedastic mixture double-autoregressive model,HMDAR)。给出了HMDAR模型的平稳性条件。利用ECM(Expectation conditional maximization)算法估计模型的参数,运用BIC(Bayes information criterion)准则选择模型.HMDAR模型分布形式的灵活性使得它能够对具有非对称或多峰分布的序列进行建模。将该模型应用于几个模拟和实际数据集均得到了较为满意的结果。特别是对于波动较大的序列,HMDAR模型能比其它模型更好的捕捉到数据序列的特征。更多还原

【Abstract】 The heteroscedastic mixture double-autoregressive (HMDAR) model was discussed. The stationary conditions were derived. The estimation of parameters was easily performed via expectation conditional maximization (ECM) algorithm. The Bayes information criterion ( BIC ) was used to select the model. The shape changing feature of conditional distributions makes the HMDAR model capable of modeling time series with asymmetric or multimodal distribution. The model was applied to several simulated and real datasets with satisfactory results. Especially to a high volatile time series, the HMDAR model appears to capture features of the data better than other competing models do.更多还原