Abstract Title

A Statistical Learning Model Utilized to Validate a Market Hypothesis

Institution

Embry-Riddle Aeronautical University

Abstract

In finance, regression models have been frequently utilized to predict the value of an asset based on its underlying traits. In prior work we built a regression model to predict the value of the S&P 500 based on macroeconomic which were selected through a process of general subjective knowledge followed by model optimization. In the present work the method of statistical machine learning is utilized to decide what predictors are to be used within the model. In addition, a well known market hypothesis “the 5 year moving average death cross” is mathematical validated and a scheme to relate those critical time periods to particular values of the regression predictors is outlined. In addition, closing comments are made to address future research ideas and how these cyclical market patterns could be related to a nonlinear PDE, most likely hyperbolic, and how modern machine learning evolutionary nonlinear fits could be utilized to discover such a PDE.

Share

COinS
 

A Statistical Learning Model Utilized to Validate a Market Hypothesis

In finance, regression models have been frequently utilized to predict the value of an asset based on its underlying traits. In prior work we built a regression model to predict the value of the S&P 500 based on macroeconomic which were selected through a process of general subjective knowledge followed by model optimization. In the present work the method of statistical machine learning is utilized to decide what predictors are to be used within the model. In addition, a well known market hypothesis “the 5 year moving average death cross” is mathematical validated and a scheme to relate those critical time periods to particular values of the regression predictors is outlined. In addition, closing comments are made to address future research ideas and how these cyclical market patterns could be related to a nonlinear PDE, most likely hyperbolic, and how modern machine learning evolutionary nonlinear fits could be utilized to discover such a PDE.