Statistical modeling based on multiple time series with various time resolution

在本研究项目中，我们开发了一种统计建模方法，用于处理以各种分辨率或不同观测间隔观测的多个时间序列。特别是，我们感兴趣的一些插值方法，利用不同的观测频率的时间序列之间的相互关系。在应用中，我们专注于在密切相关的季度时间序列（例如季度GDP的快速估计）的帮助下对年度时间序列（例如GDP的最终估计）进行去卷积。

In this research project, we developed a method of statistical modeling to work with multiple time series observed with various resolutions or different observation intervals. Especially, we are interested in some interpolation method making use of the inter-relationship between the time series with different observation frequencies. In application, we focused on the de-convolution of annual time series(the final estimate of GDP for example) with the help of closely related quarterly time series(the quick estimates of quarterly GDP for example).