Linear Predictive Coding as an Estimator of Volatility
Abstract
Description
In this paper, we present a method of estimating the volatility of a signal that displays stochastic noise (such as a risky asset traded on an open market) utilizing Linear Predictive Coding. The main purpose is to associate volatility with a series of statistical properties that can lead us, through further investigation, toward a better understanding of structural volatility as well as to improve the quality of our current estimates.