(A) statistical linear prediction in stock price process

Abstract

In viewing stock price process mathematically we need a model. The commonest mathematical model for stock price process is nonlinear stochastic Volterra integral equation. But differently from the linear case, nonlinear stochastic integral equations have few chances to obtain exact solutions. Hence for application it is general that prediction is made by approximating through numerical methods such as binomial scheme, finite difference method, Euler type iteration, Monte Carlo simulation, etc. In this thesis, we obtained the higher rate of convergence than the classical Euler type iteration method by linear prediction using statistical interpolation method.