Primjena grafova u klasifikaciji i grupiranju dokumenata

undergraduate thesis

• Title (english): Application of Graphs in Document Classification and Grouping
• Author: Leo Obadić
• Mentor: Tomislav Burić (mentor)
• Committee member: Tomislav Burić (predsjednik povjerenstva)
• Committee member: Neven Elezović (član povjerenstva)
• Committee member: Jan Šnajder (član povjerenstva)
• Granter: University of Zagreb; Faculty of Electrical Engineering and Computing; Zagreb
• Defense date and country: 2018-07-13, Croatia
• Scientific / art field, discipline and subdiscipline: TECHNICAL SCIENCES; Computing

Abstract

Cilj je ovog rada demonstrirati primjenu reprezentacije teksta grafovima kako bismo povećali preciznost klasifikacije i grupiranju teksta. U prvom dijelu objašnjava se kako tekst pretvoriti u graf. Predstavljeni su i algoritmi za izračun sličnosti grafova koji su korišteni u četiri obrađivane metode strojnog učenja (k-srednjih vrijednosti, k-reprezentanata, k-najbližih susjeda i kombinacija k-srednjih vrijednosti i k-najbližih susjeda). U drugom dijelu navedeni su eksperimentalni rezultati dobiveni u okviru programskog jezika Python.

Abstract (english)

In this paper we are trying to demonstrate application of graph models in machine learning, specifically in natural language analysis through basic clustering and classification methods. Firstly, articles gathered from few websites are preprocessed and converted to graphs and
vectors. Then, those data structures are used for clustering and classification. Methods used for clustering are k-means and k-medoids but after small modifications they are also used for classification besides k-nearest neighbours method. Furthermore, new classification method that uses both clustering and k-nearest neighbours method was proposed. Finally, after testing, graphs showed better results due to lower information loss after text to graph conversion.

• Keywords: strojno učenje; klasifikacija; grupiranje; najveći zajednički podgraf; k-srednjih vrijednosti; k-reprezentanata; k-najbližih susjeda; obrada prirodnog jezika
• Keywords (english): machine learning; classification; clustering; maximal common subgraph; k-medoids; k-means; k-NN; natural language processing
• Language: croatian
• URN:NBN: urn:nbn:hr:168:997993
• Study programme: Title: Computing; Study programme type: university; Study level: undergraduate; Academic / professional title: sveučilišni/a prvostupnik/ prvostupnica (baccalaureus/baccalaurea) inženjer/inženjerka računarstva (sveučilišni/a prvostupnik/ prvostupnica (baccalaureus/baccalaurea) inženjer/inženjerka računarstva)
• Type of resource: Text
• File origin: Born digital
• Access conditions: Closed access
• Created on: 2019-03-13 23:28:09