美股價格之高頻因果網路分析

Information Network analysis of US high frequency stock price by Causal decomposition

溫禹昊(Yu-Hao Wen)

指導教授 ： 高竹嵐

國立交通大學/理學院/統計學研究所/碩士(2020年)

摘要

在時間序列的因果關係裡面，通常是以簡單的線性模型為基準來做檢定，但我們都知道複雜系統的運行往往是非線性在運作的，這些一般描述因果關係的方法可能會低估序列間的共時性(Simultaneous)、互惠的因果相互作用性質。在本篇，我們用因果網路來描述資訊的傳遞，其中的因果聯繫不是基於預測方法，而是由價格間基於因果關係的協變以及「因在，果在，因亡，果亡」的準則定義。使用經驗模式分解，我們發現因果交互作用已經被刻進瞬時相位依賴性在特定的時間尺度下，而且這個相位相依性在與之相關的內在成分從果序列中移除後便減少。再來我們將之運用到美股那斯達克交易所1分鐘的高頻交易資料，發現每檔股票在高頻率上都有交互的因果關係。最後我們把這些關係繪製成有向網路圖並發現其因果流向的一致性，為往後研究者提供了一個參考方法。

關鍵字

經驗模態分解法 ； 本質分量函數 ； 相位相干性 ； 因果分解 ； 因果網路 ； 網路分析

並列摘要

In the causality analysis of time series, we usually do hypothesis testing based on simple linear model. But we all know complex system usually operate in non-linear mode. These predictions may underestimate the simultaneous and reciprocal nature of causal interactions. In this article, we describe information delivering by directed causal network, whose causal connection is not based on prediction, but defined by the coherence between prices and the criteria “Cause is that which put, the effect follows; and removed, the effect is removed”.
Using empirical mode decomposition, we decompose time series into several meaningful intrinsic mode functions, and show that causal interaction encoded in instantaneous phase dependency is diminished when the causal-related intrinsic component is removed from the effect. Then we utilize the method into high frequency trading data of NASDAQ, discover the causal interaction in each pair of data. Finally, we draw these connections into directed graph and find the consistency of causal flow, offing later researchers a new approach to refer.

並列關鍵字

empirical mode decomposition ； intrinsic mode function ； phase coherence ； causal decomposition ； causal network ； network analysis

參考文獻

3. Kim, Minjun, and Hiroki Sayama. “Predicting stock market movements using network science: an information theoretic approach.” Applied network science vol. 2,1 (2017): 35. doi:10.1007/s41109-017-0055-y

4. Yang, A.C., Peng, C. & Huang, N.E. Causal decomposition in the mutual causation system. Nat Commun 9, 3378 (2018).https://doi.org/10.1038/s41467-018-05845-7

1. Galilei, G. "On Motion and on Mechanics" (The University of Wisconsin Press, Madison, 1960)

2. Sun, Wenyue, Chuan Tian and Guang Yang. "Network Analysis of the Stock Market". (2015).