Markov Models Applications in Natural Language Processing: A Survey

Abstract

Markov models are one of the widely used techniques in machine learning to process natural language. Markov Chains and Hidden Markov Models are stochastic techniques employed for modeling systems that are dynamic and where the future state relies on the current state. The Markov chain, which generates a sequence of words to create a complete sentence, is frequently used in generating natural language. The hidden Markov model is employed in named-entity recognition and the tagging of parts of speech, which tries to predict hidden tags based on observed words. This paper reviews Markov models' use in three applications of natural language processing (NLP): natural language generation, named-entity recognition, and parts of speech tagging. Nowadays, researchers try to reduce dependence on lexicon or annotation tasks in NLP. In this paper, we have focused on Markov Models as a stochastic approach to process NLP. A literature review was conducted to summarize research attempts with focusing on methods/techniques that used Markov Models to process NLP, their advantages, and disadvantages. Most NLP research studies apply supervised models with the improvement of using Markov models to decrease the dependency on annotation tasks. Some others employed unsupervised solutions for reducing dependence on a lexicon or labeled datasets.

Defence against Side-Channel Attacks for Encrypted Network Communication Using Multiple Paths

“…HMMs are a widely used solution for modeling the characteristics of human language [9]. As such, they have also proven effective at modeling (and therefore analyzing) traffic with behavior derived from natural language, such as real-time Voice over IP (VoIP) calls.…”

Error Pattern Discovery in Spellchecking Using Multi-Class Confusion Matrix Analysis for the Croatian Language

“…In the context of NLP, Almutir and Nadeem [4] use confusion matrices to evaluate the performance of named-entity recognition systems by analyzing the discrepancies between predicted and actual entity labels. Pienaar and Snyman use them for the identification of eleven official South African languages [5].…”

Research on the correlation of building mechanical and electrical installation material names based on HMM-NF

“…In various fields, especially in text [3] , speech [4] , handwriting [5] and natural language [6] [7] recognition processing, Hidden Markov Models (HMM) have proven to be effective tools for sequence recognition. Nevertheless, research on applying HMM to the specific task of recognizing and classifying names of construction M&E installation materials remains sparse.…”