Title: Financial market switching-points and economic anomalies: Evidence from S\&P100
Authors: Yuan Yuan - National Institute of Informatics (Japan)
Takayuki Mizuno - National Institute of Informatics (Japan) [presenting]
Abstract: The aim is to clarify the statistical relationship between the switching points of financial markets and various economic indicators by using machine learning. Using an extensive set of macroeconomic variables, LASSO regression is applied to select the important variables which have important impact on switching-points, then the switching-points over the period 2002-2016 of S\&P100 is predicted by using the selected variables. The results show that selected variables have good out-of-sample predictive power. Moreover, ridge regression is also applied to analyze the switching-points. The contribution is to both anomaly detection of financial markets, and the application high dimensional data and machine learning to financial market.