Computing continuous-time growth models with boundary conditions via wavelets

Abstract

This paper presents an algorithm for approximating the solution of deterministic/stochastic continuous-time growth models based on the Euler's equation and the transversality conditions. The main issue for computing these models is to deal efficiently with the boundary conditions associated. This approach is a wavelets-collocation method derived from the finite-iterative trapezoidal approach. Illustrative examples are given. JEL classification codes: C63.

Wavelet-based option pricing: An empirical study

“…Ortiz-Gracia and Oosterlee (2016) use Shannon wavelet to price European options. Additional references include Bayraktar et al (2004), Hong and Kao (2004), Dong and He (2007), Esteban-Bravo and Vidal-Sanz (2007), and Haven et al (2009).…”

Forecasting industrial production in Brazil: Evidence from a wavelet approach

“…He decomposes the series via wavelets, and estimate a regression aiming to investigate whether the long-term trend or the level details present informational content. The results indicate that the long-trend and some of the detail series are significant, but this predictive power holds only for short horizons (from one to three quarters), and there is parameter instability over the sample (see also Crowley, 2005;Esteban-Bravo & Vidal-Sanz, 2007;Ramsey & Lampart, 1998;Zagaglia, 2006).…”

“…This methodology has recently received great acceptance in the financial literature (see Bayraktar et al, 2004;DiSario, Saraoglu, McCarthy, & Li, 2008;Esteban-Bravo & Vidal-Sanz, 2007;Fleming et For a multiresolution analysis, a few conditions must be satisfied: let L 2 (R) denote the space of square-integrable functions. Now consider a sequence of closed subspaces fW k g 1 k¼n (relative to the detail spaces of the series) and V n (relative to the approximation of the series) of L 2 (R), such that V n & V n+1 and \ n V n = {0}, and [ n V n = L 2 (R), which indicates that all integrable functions should be included at the highest resolution.…”

A neural network enhanced volatility component model

“…Our first contribution is that methodologically we adopt powerful computational methods in developing a novel twocomponent volatility model. The long-run component is extracted from the realized volatility via the wavelet transform, a popular de-noising and approximation method in engineering, which has found its way to the economics and finance literature (see Esteban-Bravo and Vidal-Sanz, 2007;Haven et al, 2009, for example). The long-term component is then approximated by an artificial neural network due to its strong capability in capturing trend in the time series forecasting (Zhang and Qi, 2005).…”