diplomsko delo
Ingrid Mirnik (Author), Sašo Karakatič (Mentor)

Abstract

V zaključnem delu se ukvarjamo z razvojem delno nadzorovanega meta klasifikatorja in njegovim delovanjem. Namen zaključenega dela je predstaviti koristnost delno nadzorovane klasifikacije ter uporabo te na praktičnem primeru. Rešitev smo razvili s pomočjo programskega jezika Python in scikit-learn knjižnice. Pri preverjanju delovanja klasifikatorja smo se omejili na tri različne podatkovne množice, katerim se deleži označenih podatkov spreminjajo glede na test. Primerjali smo rezultate nadzorovanih in delno nadzorovanih klasifikatorjev, ki so se vrnili podobni. Ugotovili smo, da med rezultati nadzorovanih in delno nadzorovanih klasifikatorjev ni bistvene razlike, razen v časovni zahtevnosti, ki je občutno večja pri delno nadzorovanih klasifikatorjih.

Keywords

Python;delno nadzorovana klasifikacija;strojno učenje;diplomske naloge;

Data

Language: Slovenian
Year of publishing:
Typology: 2.11 - Undergraduate Thesis
Organization: UM FERI - Faculty of Electrical Engineering and Computer Science
Publisher: [I. Mirnik]
UDC: 004.434:004.8(043.2)
COBISS: 93977859 Link will open in a new window
Views: 267
Downloads: 40
Average score: 0 (0 votes)
Metadata: JSON JSON-RDF JSON-LD TURTLE N-TRIPLES XML RDFA MICRODATA DC-XML DC-RDF RDF

Other data

Secondary language: English
Secondary title: Semi-supervised meta classifier in the Python programming language
Secondary abstract: In this thesis, we are dealing with the development of a semi-supervised meta classifier and its operation. The purpose is to present the usefulness of semi-supervised classification and its usage on a practical example. We developed the solution using Python programming language and the scikit-learn library. When testing the classifier, we limited ourselves to three distinctive data sets, which labeled data proportions change with different tests. We compared the results of supervised and semi-supervised classifiers, which came back similar. We disclosed no notable difference between the results of supervised and semi-supervised classifiers, except the time complexity, which is significantly higher for semi-supervised classifiers.
Secondary keywords: Python;semi-supervised classification;machine learning;
Type (COBISS): Bachelor thesis/paper
Thesis comment: Univ. v Mariboru, Fak. za elektrotehniko, računalništvo in informatiko, Informatika in tehnologije komuniciranja
Pages: VII, 56 str.
ID: 13305815