Combining IR Models for Bengali Information Retrieval

Combining IR Models for Bengali Information Retrieval

Soma Chatterjee (Jadavpur University, Kolkata, India) and Kamal Sarkar (Computer Science and Engineering, Jadavpur University, Kolkata, India)
Copyright: © 2018 |Pages: 16
DOI: 10.4018/IJIRR.2018070105

Abstract

Word mismatch between queries and documents is a fundamental problem in information retrieval domain. In this article, the authors present an effective approach to Bengali information retrieval that combines two IR models to tackle the word mismatch problem in Bengali IR. The proposed hybrid model combines the traditional word-based IR model with another IR model that uses semantic text similarity measure based on vector embeddings of words. Experimental results show that the performance of our proposed hybrid Bengali IR model significantly improves over the baseline IR model.
Article Preview
Top

Previous Work

The most common information retrieval models that are in use are Boolean retrieval model, Probabilistic model, Vector Space Model and Language model.

The earliest works on Information retrieval model was devoted to Boolean Retrieval Model which is the simplest and most widely used model. The model relies simply on Boolean operators like AND, OR and NOT. The terms are linked together with these operators. However, it searches the exact words thus a user has to have some idea about the query he is using, for example misspelling will not give an intended result. The very common problem of Boolean retrieval model is term mismatch problem (Zhao, 2012). To deal with this, the common strategy is to use Stemming which reduces a term to its morphological stem and using it as a prefix, users can retrieve many terms that are conceptually related to the original term (Marcus, 1991).

Complete Article List

Search this Journal:
Reset
Open Access Articles: Forthcoming
Volume 10: 4 Issues (2020): 1 Released, 3 Forthcoming
Volume 9: 4 Issues (2019)
Volume 8: 4 Issues (2018)
Volume 7: 4 Issues (2017)
Volume 6: 4 Issues (2016)
Volume 5: 4 Issues (2015)
Volume 4: 4 Issues (2014)
Volume 3: 4 Issues (2013)
Volume 2: 4 Issues (2012)
Volume 1: 4 Issues (2011)
View Complete Journal Contents Listing