A generalized Erdős-Rényi law for sequence analysis problems (Q5926525)

This research is part of the author's (1998) Ph. D. thesis at the University of Boulder, Colorado. Main scope is to characterize a general class of random variables, including scoring functions arising in computational molecular biology, that obey an Erdős-Rényi type law of large numbers under certain conditions. A number of sequence analysis problems which have previously been studied is encompassed by this class, but also two new laws of large numbers follow from an application of this generalized Erdős-Rényi theorem. A stochastic subadditive property is key to the proof of the general theorem.