Best Subsets

The best subsets search method can be based on three different test statistics: the score statistic, the model likelihood, and the AIC. Note that, since the score statistic does not require iterative computations, best subset selection based on the score statistic is computationally fastest, while selection based on the other two statistics usually provides more accurate results.

In general, a model containing p possible predictors has 2p-1 possible subsets of predictors available for model consideration. In the best subsets method, one of the aforementioned three test statistics is computed for every possible predictor subset and a spreadsheet is generated such that the subsets are sorted from best to worst by test statistic. Due to the number of possible subsets to consider as they grow exponentially with increasing number of effects, Statistica® only performs a best subset search when the number of effects in the model is 13 or fewer.