Statistical Inference on Long-Memory Time Series and its Applications to Economic Data

在本研究中，我们提出了各种长记忆时间序列的半参数统计推断，如长记忆信号加噪声过程和周期性长记忆时间序列。将该方法应用于一个长记忆随机波动模型，考察了汇率波动的长记忆性。我们还对日本工业生产指数的增长率进行了实证分析，并检测了其周期性持续性。

In this research, we proposed semiparametric statistical inference on various long-memory time series, such as long-memory signal plus noise processes and cyclical long-memory time series. By applying the proposed method to a long-memory stochastic volatility model, we examined long-memory in the volatility of exchange rates. We also provided an empirical analysis of the growth rate of Japan's industrial production index and detected its cyclical persistence.