Syntactic Information Retrieval

Abstract

Natural language processing (NLP) techniques are believed to have the potential to aid information retrieval (IR) in terms of retrieval accuracy. In this paper we report a proof of concept study on a new approach to NLP-based IR that we propose. Documents and queries are represented as syntactic parse trees, which are generated by a natural language parser. Based on this tree structured representation of documents and queries, the matching between a document and a query is executed on their tree representations, with tree comparison as the key operation. An IR experiment is designed to test if this approach is feasible. Experimental results show that this approach is promising and has the potential to outperform the standard bag of words approach to information retrieval, especially in response to long queries.

Pruning dependency trees in a syntactic information retrieval model

“…Once this tree structured representation of documents and query is built, the matching between a document and a query is executed directly on the tree representation level, with the method of comparing trees as the key operation. The experiment carried out in [1] proves that SIR is feasible and performs better than the vector space model especially in response to long queries.…”

“…In this paper we present new research work that goes further in exploring and examining the novel IR model (SIR) we proposed in [1]. From a syntactic theoretical perspective, we investigate the necessaries of pruning raw parse tree in SIR and propose a pruning strategy based on a consideration of the role of word in the dependency tree and their part of speech.…”

“…An IR experiment is designed to evaluate SIR on a TREC dataset of full documents, which represent a considerable scaling of our previous work [1]. At the same time, the same IR experiment is executed on a well known term based IR system, Terrier [12].…”

“…In [1], a new IR strategy (SIR) was proposed with NLP techniques involved at the syntactic level. Within SIR, documents and the query are represented on the basis of a syntactic data structure of the natural language text -syntactic parse tree, in which syntactic relationships between words are identified and represented in the form of a tree.…”

Design and Implementation of Virtual Logistics Storage System

“…Seen from Fig. 1, system framework adopts improved MVC (model View Controller) design thought [2], i.e., an input, processing, output of application are separated through model, view and controller individually. In this process, the view is be on behalf of the user interface, the model accept view request data, and return the final processing results, the controller can match model and view together by receiving request from the users and finish the user's request together.…”