.png){kind=icon}
.png){kind=icon}
Return
A high-performance FAQ retrieval method using minimal differentiator expressions
.png){kind=icon}
PRE
AI
DOI：10.1016/j.knosys.2012.05.015.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
Case-Based Reasoning (CBR) has proven to be a very useful technique to solve problems in Closed-Domains Question Answering such as FAQ retrieval. Instead of trying to uderstand the question this method consists of retrieving the most similar case (Question/Answer pairs) among all cases by analogy. Keyword comparison criterion or statistical approaches are often used to implement similarity measure. However, those methods present the following disadvantages. On the one side, choosing keywords is an expert-knowledge domain-dependant task that is often performed manually. Furthermore, keyword comparison criterion does not guarantee the total differentiation among cases. On the other side, statistical approaches do not perform with enough information in sentence-level problems and are not interpretable. In order to alleviate these deficiencies we present a new method called the Minimal Differentiator Expressions (MDE) algorithm. This algorithm automatically obtains a set of linguistic patterns (expressions) used to retrieve the most relevant case to the user question. Those patterns present the following advantages: they are composed by the simplest sets of words which permit differentiation among cases and they are easily interpretable. (C) 2012 Elsevier B.V. All rights reserved.
Keywords:
Question answering
FAQ retrieval
Question recognition
CBR
Natural language
AI Summary
Key information extracted from the uploaded paper, including a brief overview, abstract, background, key highlights, visual analysis, and future outlook.