The purpose of this paper is to associate a probabilistic confidence with the stock market interval forecasts obtained with the interval least squares (ILS) algorithm. The term probabilistic confidence in this paper means the probability of a point observation that will fall in the interval forecast.
Using confidence interval as input, annual ILS forecasts of the stock market were made. Then the probability of point observation that fall in the annual forecasts was examined empirically.
When using confidence interval as ILS input, the stock market annual interval forecasts may have the same level of confidence as that of the input intervals.
At the same confidence level, the ILS can produce much better quality forecasts than the traditional ordinary least squares method for the stock market. Although the algorithmic approach can be applied to analyze other datasets, one should examine implications of computational results as always.
Results of this specific paper may be interesting to executive officers, other financial decision makers and to investors.
Although the ILS algorithm has been recently developed in forecasting the variability of the stock market, this paper presents the first successful attempt in associating a probabilistic confidence with ILS interval forecasts.
