Using FCA and Concept Explications for Finding an Appropriate Concept

Result code in IS VaVaI

<a href="https://www.isvavai.cz/riv?ss=detail&h=RIV%2F47813059%3A19240%2F21%3AA0000920" target="_blank" >RIV/47813059:19240/21:A0000920 - isvavai.cz</a>

Alternative codes found

RIV/61989100:27240/21:10248836

Result on the web

<a href="https://nlp.fi.muni.cz/raslan/raslan21.pdf" target="_blank" >https://nlp.fi.muni.cz/raslan/raslan21.pdf</a>

DOI - Digital Object Identifier

—

Result language

angličtina

Original language name

Using FCA and Concept Explications for Finding an Appropriate Concept

Original language description

This paper introduces the method of discovering a plausible atomic concept that corresponds to the generated molecular concept explication and known attributes’ values and properties of objects falling under the concept. First, we summarize the process of concept explication via the symbolic method of supervised machine learning from formalized natural language sentences. To obtain particular concept explications, we exploit heuristic procedures that operate on the symbolic representation of current hypothesis and example to obtain particular concept explications. These explications serve as descriptions of the sought atomic concept accordingly to the given text sources. Afterwards, the method of searching for the appropriate concept based on attributes’ values is outlined. Thus, user can seek a specific concept, which can be vague or inaccurate, among the so-extracted explications. We focus on a situation in which the user knows basic properties or attributes’ values and searches for a suitable atomic concept that is described by these properties or attributes’ values. To explain the process, we summarize the creation of explications and the method of Formal Concept Analysis (FCA) as a theoretical background. As a result, we present to the user an appropriate atomic concept. The whole method is demonstrated by a few examples.

Czech name

—

Czech description

—

Type

D - Article in proceedings

CEP classification

—

OECD FORD branch

10201 - Computer sciences, information science, bioinformathics (hardware development to be 2.2, social aspect to be 5.8)

Project

—

Continuities

S - Specificky vyzkum na vysokych skolach

Publication year

2021

Confidentiality

S - Úplné a pravdivé údaje o projektu nepodléhají ochraně podle zvláštních právních předpisů

Article name in the collection

Proceedings of Recent Advances in Slavonic Natural Language Processing, RASLAN 2021,

ISBN

9788026316701

ISSN

2336-4289

e-ISSN

—

Number of pages

12

Pages from-to

49-60

Publisher name

Tribun EU

Place of publication

Brno

Event location

Karlova Studánka

Event date

Dec 10, 2021

Type of event by nationality

EUR - Evropská akce

UT code for WoS article

—