diplomsko delo
Jernej Vivod (Author), Marko Robnik Šikonja (Mentor)

Abstract

Algoritem Relief in njegove posplošitve so filtrirne metode vrednotenja atributov, ki jih odlikuje občutljivost na medatributne interakcije. Diplomsko delo pričnemo z opisom problematike izbora atributov in podamo motivacijo za uporabo algoritma Relief in njegovih posplošitev. Opišemo, po našem prepričanju, danes najpogosteje uporabljene posplošitve algoritma Relief v klasifikaciji. Predstavimo koncept naučenih metrik in podrobneje predstavimo različnost na osnovi mase ter preostale naučene metrike, ki jih uporabimo v kontekstu opisanih algoritmov. Diplomsko delo sklenemo z empiričnim vrednotenjem implementiranih algoritmov in metrik, kjer uporabimo Bayesov hierarhični korelirani t-test in izris rezultatov prečnega preverjanja za različne velikosti množic najbolje ocenjenih atributov. Na koncu izpostavimo omejitve uporabljene metodologije vrednotenja in podamo iztočnice za nadaljnje raziskovalno delo.

Keywords

strojno učenje;umetna inteligenca;vrednotenje atributov;rangiranje atributov;izbor atributov;klasifikacija;računalništvo in informatika;univerzitetni študij;diplomske naloge;

Data

Language: Slovenian
Year of publishing:
Typology: 2.11 - Undergraduate Thesis
Organization: UL FRI - Faculty of Computer and Information Science
Publisher: [J. Vivod]
UDC: 004.85(043.2)
COBISS: 1538334915 Link will open in a new window
Views: 814
Downloads: 253
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: Feature evaluation with generalizations of Relief algorithm
Secondary abstract: The Relief algorithm and its generalizations form a group of filter-based feature evaluation algorithms that are sensitive to feature interactions. We describe the problem of feature selection and present motivation for the application of Relief and its generalizations. We describe all commonly used generalizations of Relief used in classification. We describe the concept of learned metric functions and describe mass-based dissimilarity as well as other learned metric functions, studied in the context of described algorithms. We conclude the thesis with an empirical evaluation of implemented algorithms and metrics. We use the Bayesian hierarchical correlated t-test and plot cross validation results against different cardinalities of feature subsets. We analyze the limitations and assumptions of our evaluation methodology and present ideas for further research.
Secondary keywords: machine learning;artificial intelligence;feature evaluation;feature ranking;feature selection;classification;computer and information science;diploma;
Type (COBISS): Bachelor thesis/paper
Study programme: 1000468
Embargo end date (OpenAIRE): 1970-01-01
Thesis comment: Univ. v Ljubljani, Fak. za računalništvo in informatiko
Pages: 108 str.
ID: 11221284