Analysis of volatile export data of fruit and vegetable seeds: An application of stochastic volatility model using the particle filter

Financial time-series data of certain essential commodities show heteroscedasticity, andthebehaviour of prices of such commodities is fundamental to policy makers. Out of many approaches available in the literature for modelling volatile data sets, one approach is the promising methodology of Stochastic Volatility (SV) model. A heartening feature of SV is that it assumes the volatility to be an unobservable state variable following some latent stochastic process. In this paper, procedure for estimation of parameters of SV, using Particle Filter (PF), a powerful Monte Carlo technique, is thoroughly discussed and subsequently, the unobservable volatility along with the parameters of the modelis estimated. To this end, relevant computer program in Matlab software package isdeveloped. As an illustration, the month-wise total export of fruits and vegetable seeds from India are considered. Comparative study of the fitted SV model (SVPF) is also carried out with SV model fitted through Kalman filter (SVKF)by calculating various measures of goodness of fit. Furthermore, the forecasting performance is also examined using appropriate statistical measures. Finally, it is concluded that SVPF performed better than SVKF for modelling as well as forecasting the data under consideration.