本文選題：趨勢持續(xù)期　切入點：經(jīng)驗模態(tài)分解　出處：《統(tǒng)計與決策》2017年20期 　論文類型：期刊論文

【摘要】：文章針對我國滬深300股指期貨高頻數(shù)據(jù)時間序列具有趨勢運動特性,提出了趨勢持續(xù)期模型。首先采用泊松過程對趨勢持續(xù)期的市場微觀結(jié)構(gòu)進行建模,得出了趨勢持續(xù)期在理論上服從Gamma分布;基于經(jīng)驗模態(tài)分解算法提取股指期貨日內(nèi)高頻交易數(shù)據(jù)的趨勢持續(xù)期,采用最大似然估計法,估計趨勢持續(xù)期的Gamma分布參數(shù),同時通過Kolmogorov-Smirnov檢驗驗證了模型的有效性;最后對不同采樣間隔下的趨勢持續(xù)期進行標準化處理,趨勢持續(xù)期模型具有很好的穩(wěn)健性。
[Abstract]:In this paper, a trend duration model is proposed for the time series of high frequency data of CSI 300 stock index futures. Firstly, Poisson process is used to model the market microstructure of the trend duration. Based on empirical mode decomposition algorithm, the trend duration of intraday high frequency trading data of stock index futures is extracted, and the Gamma distribution parameters of trend duration are estimated by using maximum likelihood estimation method. At the same time, the validity of the model is verified by Kolmogorov-Smirnov test. Finally, the trend duration model with different sampling intervals is standardized, and the trend duration model has good robustness.
【作者單位】： 北京大學經(jīng)濟學院;
【分類號】：F224;F724.5
，

本文編號：1558602