Klasifikacija uzoraka predočenih vektorima jezičnih varijabli

scientific master's theses

• Author: Jasna Drozdek
• Mentor: Slobodan Ribarić (mentor)
• Committee member: Bojana Dalbelo-Bašić (predsjednik povjerenstva)
• Committee member: Slobodan Ribarić (član povjerenstva)
• Committee member: Nikola Pavešić (član povjerenstva)
• Granter: University of Zagreb; Faculty of Electrical Engineering and Computing; (Department of Electronics, Microelectronics, Computer and Intelligent Systems); Zagreb
• Defense date and country: 2000-10-24, Croatia
• Scientific / art field, discipline and subdiscipline: TECHNICAL SCIENCES; Computing
• Universal decimal classification (UDC): 004 - Computer science and technology. Computing. Data processing

Abstract

Tema magistarskog rada je klasifikacija uzoraka predočenih vektorima jezičnih varijabli. Prvo se definiraju pojmovi neizrazitog skupa i neizrazitog broja, te se opisuje pravilo proširenja algebarskih opracija sa skupa realnih na skup neizrazitih brojeva. Zatim se uvodi definicija jezične varijable, te se opisuju Booleove i strukturne jezične varijable. Nakon toga definira se ndimenzionalni prostor jezičnih varijabli, uvodi se (pseudo)metrika na tom prostoru i navodi se algoritam normalizacije jezičnih varijabli. Formule za opisane (pseudo)metrike su zatim prilagođene za trapezoidalne jezične varijable. Opisuju se crisp q-NN (q najbližih susjeda) i qNP (q najbližih protoipova) algoritmi klasifikacije uzoraka, kao i njihove neizrazite verzije. Neizraziti q-NN i q-NP algoritmi su zatim prilagođeni za klasifikaciju uzoraka iz ndimenzionalnog prostora jezičnih varijabli. Na kraju je dana usporedba neizrazitih q-NN i q-NP metoda klasifikacije uzoraka koji su vektori jezičnih varijabli s klasifikatorom A. K. Natha i T. T. Leea. U dodatku A je opisana programska dokumentacija za program koji obavlja klasifikaciju uzoraka iz n-dimenzionalnog prostora jezičnih varijabli neizrazitim q-NN i q-NP algoritmima, koji je priložen na CD-ROMu.

Abstract (english)

The theme of this master’s thesis is pattern classification for patterns which are vectors of linguistic variables. Firstly, fuzzy set and fuzzy number are defined, and the extension principle for algebraic operations on the set of fuzzy numbers is given. The definition of linguistic variable is introduced, and special types of linguistic variables – Boolean and structured ones are described. Secondly, the n-dimensional space of linguistic variables is defined and (pseudo)metric functions on that space are introduced. The normalization algorithm for linguistic variables is given. Formulas for (pseudo)metric functions are applied in n-dimensional space of trapezoidal linguistic variables. Crisp q-NN (nearest neighbor) and q-NP (nearest prototype) classification algorithms are described and their fuzzy versions are given. Those fuzzy q-NN and q-NP algorithms are adopted for classification of the patterns which are vectors of linguistic variables. Finally, classifier of A. K. Nath and T. T. Lee is compared with fuzzy q-NN and q-NP classification algorithms for patterns that are vectors of linguistic variables. In the Appendix A program documentation for program that classifies n-dimensional vectors of linguistic variables using fuzzy q-NN and q-NP classification algorithms is given, and this program is recorded on CD-ROM and added to the thesis.

• Keywords: Neizraziti skup; neizraziti broj; pravilo proširenja; jezična varijabla; ndimenzionalni prostor; metrika; normalizacija; uzorak; klasifikacija; najbliži susjed
• Keywords (english): Fuzzy set; fuzzy number; extension principle; linguistic variable; n-dimensional space; (pseudo)metric function; normalization; pattern; classification; nearest neighbor
• Language: croatian
• URN:NBN: urn:nbn:hr:168:876857
• Study programme: Title: Postgraduate master programe in computing; Study programme type: university; Study level: postgraduate; Academic / professional title: magistar inženjer računarstva (magistar inženjer računarstva)
• Type of resource: Text
• File origin: Born digital
• Access conditions: Closed access
• Created on: 2019-04-10 13:02:16