Abstract:
Markov chain theory is a popular statistical tool in applied probability that is quite useful in modelling
real-world computing applications. Over the past years; there has been grown interest to employ Markov
chain theory in statistical learning of temporal (i.e. time series) data. A wide range of applications found to
utilize Markov concepts; such applications include computational linguists, image processing,
communications, bioinformatics, finance systems, etc .In fact, Markov processes based research applied
with great success in many of the most efficient natural language processing (NLP) tools. Hence, this paper
explores the Markov chain theory and its extension hidden Markov models (HMM) in (NLP) applications.
This paper also presents some aspects related to Markov chains and HMM such as creating transition and
observation matrices, calculating data sequence probabilities, extracting the hidden states, and profile
HMM.