First thoughts on determining a method for fast autocorrelation classification

Classification of Boolean functions is a useful tool; it greatly reduces the 2²ⁿ Boolean functions to a much more feasible number. Recent work has suggested the use of the autocorrelation transform to generate coefficients that may be used as a classification tool for Boolean functions. An outstanding question, however, is how to quickly identify whether or not two given functions are in the same class, given that generating the 2ⁿ-sized spectrum may not always be feasible. This work addresses this question, and presents some preliminary heuristics that are analyzed for their suitability in solving this problem.