Estimating the number of states of a finite-state source
J. Ziv, Neri Merhav · IEEE Transactions on Information Theory · 1992
The problem of estimating the number of states of a finite-alphabet, finite-state source is investigated. An estimator is developed that asymptotically attains the minimum probability of understanding the number of states, among all estimators with a prescribed exponential decay rate of overestimation probability. The proposed estimator relies on the Lempel-Ziv data compression algorithm in an intuitively appealing manner.>