SOME METHODOLOGICAL ASPECTS RELATED TO THE PREDICTION OF LARGE AFTERSHOCKS