A SYSTEM FOR INTEGRATION OF RELATIONAL DATABASE AND SIMPLE SENTENCES IN NATURAL LANGUAGE USING THE 'NODE OF KNOWLEDGE' CONCEPTUAL FRAMEWORK
Autor: | Ašenbrener Katić, Martina |
---|---|
Přispěvatelé: | Pavlić, Mile, Radovan, Mario, Poščić, Patrizia, Jakupović, Alen |
Jazyk: | chorvatština |
Rok vydání: | 2018 |
Předmět: |
relational database
knowledge representation rules of transformation information systems entity and relationship method |
Popis: | Postojeći informacijski sustavi i njihove pripadne baze podataka ne mogu prihvatiti znanja zapisana u rečenicama prirodnog jezika. Na osnovi konceptualnog okvira "Node of Knowledge (NOK)", analizira se i modelira novi sustav temeljen na NOK metodi koji će u svojoj bazi podataka (BPNOK) čuvati tekstualno znanje iz rečenica prirodnog jezika bez gubitka semantike. Osim toga, sustav temeljen na NOK metodi može prihvatiti i pitanja dana upitnim rečenicama prirodnog jezika te na temelju znanja pohranjenog u BPNOK dati odgovor na postavljeno pitanje. Cilj ovog rada je obogatiti postojeće informacijske sustave i njihove baze podataka sustavom za prihvat i predstavljanje znanja zapisanih u tekstovima s jednostavnim rečenicama prirodnog jezika, odnosno sustavom temeljenom na konceptualnom okviru "Node of Knowledge". Integracija dvaju sustava omogućuje dodavanje novih znanja vezanih za podatke u bazi podataka postojećeg informacijskog sustava bez izmjene postojeće sheme baze podataka. Rezultat tog proširenja je integralna semantički bogatija baza podataka. U radu je definiran konceptualni okvir "Node of Knowledge". Definirana su pravila transformacije izjavnih i upitnih rečenica prirodnog jezika u FNOK, odnosno QFNOK zapis. Također, definirani su modeli sustava temeljenog na NOK metodi, algoritmi za transformaciju FNOK i QFNOK zapisa u relacijsku bazu podataka te algoritam za kreiranje odgovora. Razvijen je model integrirane baze podataka, odnosno baze podataka postojećeg informacijskog sustava koji je integriran sa sustavom koji omogućuje unos znanja iskazanog rečenicama prirodnog jezika. Sada je moguće definirati pitanja prirodnog jezika koja obuhvaćaju znanje iz oba sustava. Upiti su vezani za obje relacijske baze podataka. Osim toga, omogućeno je i dobivanje odgovora na postavljeno pitanje. The existing information systems and their associated databases cannot accept the knowledge written in sentences in natural language. Based on the conceptual framework Node of Knowledge (NOK), a new system based on the NOK method is analysed and modelled. It will save textual knowledge from natural language sentences into its database (DBNOK) without loss of semantics. In addition, a system based on NOK method can also accept questions in the form of interrogative sentences of natural language, and based on the knowledge stored in DBNOK, provide answers to the questions. The aim of this paper is to enrich existing information systems and their databases with a system for the reception and representation of knowledge written in texts using simple sentences of natural language; i.e. with a system that is based on the conceptual framework of "Node of Knowledge". The integration of two systems allows the addition of new knowledge related to the data in the database of the existing information system without changing the existing database schema. The result of this expansion is an integral, semantically richer database. The paper defines the conceptual framework "Node of Knowledge". Rules for transformation of declarative and interrogative sentences in natural language into FNOK and QFNOK records are defined. Also, models of the system based on NOK method, algorithms for transformation of FNOK and QFNOK records into a relational database and an algorithm for creating answers are defined. A model of an integrated database, i.e. a database of the existing information system that is integrated with a system which allows entering knowledge expressed in sentences in natural language is developed. It is now possible to define questions in natural language which include knowledge of both systems. Queries are connected to both databases. In addition, it is possible to obtain answers to the set questions. |
Databáze: | OpenAIRE |
Externí odkaz: |