A review and a synthesis of the fast Fourier transform algorithms for exact analysis of discrete data

摘要: Abstract The fast Fourier transform (FFT) has been used to devise efficient algorithms in four types of discrete data problems. These are exact inference on multinomial data, several 2 × tables, unstratified and stratified linear logistic models, multi-way contingency table data. We review these applications. Normally, FFT presented terms characteristic functions; we do that polynomials. A polynomial-based approach is simpler derives naturally from the fact distributions for many including those reviewed, arise This approach, moreover, facilitates a synthesis diverse field inference.