An Integrated Approach For Stock Price Forecasting

This article faces the problem of stock price forecasting based on an integrated approach in which the modeling of high frequency financial data (duration, volume and bid-ask spread) uses a contemporaneous ordered probit model – the price changes (measured in numbers of ticks) are the interest variable. Here, the formulation introduced by Raposo and Veiga (2004) – EMACM – is used in order to capture the dynamic that high frequency variables present, and its forecasting function is taken as proxy to the contemporaneous information necessary to the price model being created. In that context, the main purpose of the article is to compare the performance of the current model against NAIVE, testing the use of real data instead of the results of the forecasting function.