Short Term Forecasting of Stock Market Share Prices using Arima Models

Abstract

The normal theory tests for the equality of k variances and co¬variance matrices are described. The size of the likelihood ratio test both in univariate and mulii-variate case is obtained. It is shown that the size of the test is influenced by kurtosis when the parent distribution is non-normal: and it is also influenced by the increase in groups (k)and dimensions (p).