N-Gram Based Semantic Enhanced Model For Product Information Retrieval

Abstract

The current information retrieval mechanisms are
based on models such as Boolean model, extended
Boolean model, vector space model, and probabilistic model
and language models. However, these models fall short of
expectations, leading to misunderstanding of the user query
and therefore the information retrieved fail to meet user
expectations. In this paper, a novel search technique is
proposed as the possible solution to the problems inherent in
the current information retrieval models. To achieve this
objective, an experimental research design was utilized. This
new technique is based on the concept of N-gram coupled with
a conceptual search in which information is searched b a s ed
on meaning instead of matching of keywords. Thereafter, Ngrams
are employed to tokenize and display the retrieved
information. To validate the proposed approach, a number of
experimentations were carried out based on the current search
criteria as well as the N-gram based semantic search. The
results obtained demonstrated that the proposed information
retrieval technique outperformed the current models in terms
of precision, recall and F-measure. As such, this model
proves significant in the information retrieval process as it
accomplishes search by meaning instead of keyword based
searching.