The *Tsallis entropy* of a continuous random variable ''X'' is defined by

''H_q(X) = (1 / (q - 1))(1 - int_{RR^n} f^q(x) dx)''

where ''f : RR^n -> RR'' is the probability density function of ''X'', and ''q in RR''. As ''q'' approaches 1, the Tsallis entropy approaches the Shannon differential entropy. This limit can be made part of the definition to make Tsallis entropy continuous on ''q''.

Analytic solutions for Tsallis entropies

Leonenko-Pronzato-Savani estimator