Analisi asintotica per un Wright-Fisher Hidden Markov Models

In recent years, there has been an increasing interest in models for time series data which are both flexible and explainable. A remarkable example is a Hidden Markov Model with Wright-Fisher process as latent signal, together with its Nonparametric extension. In this work we analyse the asymptotic performances of the Hidden Markov Model with a latent Wright-Fisher signal in the setting of independent and identically distributed data. Our analysis shows that, for several estimators that arise in such model, conditioning to the data and conditioning to the related emitting distributions are asymptotically equivalent, as the amount of data diverges.