Reference Hub3
A New Approach Using Hidden Markov Model and Bayesian Method for Estimate of Word Types in Text Mining

A New Approach Using Hidden Markov Model and Bayesian Method for Estimate of Word Types in Text Mining

Adem Doganer, Sinan Calik
Copyright: © 2017 |Volume: 8 |Issue: 4 |Pages: 13
ISSN: 1947-8208|EISSN: 1947-8216|EISBN13: 9781522514145|DOI: 10.4018/IJKSS.2017100102
Cite Article Cite Article

MLA

Doganer, Adem, and Sinan Calik. "A New Approach Using Hidden Markov Model and Bayesian Method for Estimate of Word Types in Text Mining." IJKSS vol.8, no.4 2017: pp.17-29. http://doi.org/10.4018/IJKSS.2017100102

APA

Doganer, A. & Calik, S. (2017). A New Approach Using Hidden Markov Model and Bayesian Method for Estimate of Word Types in Text Mining. International Journal of Knowledge and Systems Science (IJKSS), 8(4), 17-29. http://doi.org/10.4018/IJKSS.2017100102

Chicago

Doganer, Adem, and Sinan Calik. "A New Approach Using Hidden Markov Model and Bayesian Method for Estimate of Word Types in Text Mining," International Journal of Knowledge and Systems Science (IJKSS) 8, no.4: 17-29. http://doi.org/10.4018/IJKSS.2017100102

Export Reference

Mendeley
Favorite Full-Issue Download

Abstract

Determining the structure of words in the text for the operations such as automated information extraction and text summarization of the text is essential. In computers, textual analysis to define the type of the word is considered as a vital advantage. Defining the types of words provides an estimate of the sequence of words in the sentence. In this article, estimating types of Turkish words is provided by developing a Hidden Markov Model and a Bayesian-based new model. In this model, an algorithm is developed which separates the suffixes of the words and grouping the words by counts of characters that suffixes of the words receive. A text composed of 584 Turkish words is used for the testing the dependability of the model. The model has achieved a high success rate in predicting the types of Turkish words.

Request Access

You do not own this content. Please login to recommend this title to your institution's librarian or purchase it from the IGI Global bookstore.