Formalna analiza koncepata i kombinatorno testiranje za automatiziranu provjeru znanja u sustavima za e-učenje

doctoral thesis

• Title (english): Formal concept analysis and combinatorial testing for automated assessment in e-learning systems
• Author: Frano Škopljanac-Mačina
• Mentor: Bruno Blašković (mentor)
• Committee member: Gordan Ježić (predsjednik povjerenstva)
• Committee member: Mirko Randić (član povjerenstva)
• Committee member: Ivona Zakarija (član povjerenstva)
• Granter: University of Zagreb; Faculty of Electrical Engineering and Computing; (Department of Electrical Engineering Fundamentals and Measurements); Zagreb
• Defense date and country: 2023-04-26, Croatia
• Scientific / art field, discipline and subdiscipline: TECHNICAL SCIENCES; Electrical Engineering
• Universal decimal classification (UDC): 621.3 - Electrical engineering

Abstract

Tema ove disertacije je istraživanje primjene metode formalne analize koncepata i tehnike kombinatornog testiranja za automatiziranje pripreme i odabira pitanja za provjere znanja u sustavima za e-učenje. Na temelju skupa ispitnih pitanja označenih s definiranim atributima gradi se primjenom metode za formalnu analizu koncepata formalni kontekst. Potom se automatiziranom metodom kombinatornog testiranja generira gotovo minimalan broj testnih slučajeva opisanih s definiranim atributima iz formalnog konteksta tako da je svaka n-torka atributa zadane veličine n pokrivena s barem jednim testnim slučajem odnosno opisom pitanja. Nakon što se pronađe sažeti skup pitanja koji pokriva sve generirane testne slučajeve prelazi se na izgradnju konceptualne rešetke primjenom metode formalne analize koncepata. Konceptualna rešetka predstavlja ontologiju dijela nastavnog gradiva opisanog s odabranim sažetim skupom ispitnih pitanja. Konceptualna rešetka se automatski topološki sortira kako bi se iz nje dobio odgovarajući potpuno uređeni skup formalnih koncepata, a iz njega se potom automatski izdvajaju prikladni nizovi pitanja za formativnu provjeru znanja u sustavu za e-učenje. Formativna provjera se provodi kroz vlastiti sustav za e-učenje, koji studente vodi po pripremljenim nizovima pitanja i pritom nudi pomoć u obliku interaktivnih nastavnih materijala. Izgrađen je i model za verifikaciju predložene metode uz pomoć nadograđenog i automatiziranog L* algoritma i metode provjere modela s alatom Spin. Izvorni doprinosi su: 1) formalni opis metode strojnog učenja za izgradnju ontologije domenskog znanja u obliku konceptualne rešetke na temelju sažetog skupa semantički označenih ispitnih pitanja, 2) metoda za automatiziranu provjeru znanja utemeljena na automatskom odabiru nizova ispitnih pitanja iz konceptualne rešetke i 3) prototip sustava za verifikaciju metode za automatiziranu provjeru znanja utemeljen na formalnoj metodi provjere modela.

Abstract (english)

The topic of this dissertation is research on the application of Formal concept analysis method and combinatorial testing technique to automate preparation and selection of questions for assessments in e-learning systems. Based on the exam question set labelled with defined attributes a formal context is built using formal concept analysis method. Then, combinatorial testing method generates almost minimal number of test cases described with defined attributes from the formal context so that each tuple of attributes of a given size n is covered by at least one test case or a question description. After finding a concise set of questions that covers all generated test cases, we can build its conceptual lattice using the method of formal concept analysis, as an ontology of the teaching material described with a selected concise set of exam questions. The concept lattice is automatically topologically sorted to obtain a totally ordered set of formal concepts, and then appropriate sequences of questions for formative assessment are automatically extracted from it. Students can the formative assessment though our e-learning system which guides them over prepared question sequences and offers help in the form of interactive teaching materials. A model for verification of the proposed method was built using upgraded and automated L* algorithm and the model checking method with the Spin tool. Original scientific contribution of the dissertation: 1) formal description of a machine learning method for building domain knowledge ontology as a concept lattice based on a concise set of semantically labelled exam questions, 2) method for automated assessment based on automatic selection of exam questions sequences from a concept lattice and 3) prototype verification.

• Keywords: formalna analiza koncepata (FCA); kombinatorno testiranje; topološko sortiranje; formalna metoda provjere modela; Spin/Promela; L* algoritam; automatizirana provjera znanja; oblikovanje sustava za e-učenje
• Keywords (english): formal concept analysis (FCA); combinatorial testing; topological sorting; formal method of model checking; Spin/Promela; L* algorithm; automated assessment; e-learning system design
• Language: croatian
• URN:NBN: urn:nbn:hr:168:270490
• Promotion: 2023
• Study programme: Title: Doctoral study programme "Electrical Engineering and Computing"; Study programme type: university; Study level: postgraduate; Academic / professional title: doktor/doktorica znanosti, po-dručje tehničkih znanosti (doktor/doktorica znanosti, po-dručje tehničkih znanosti)
• Type of resource: Text
• Extent: 206 str. : graf. prikazi ; 31 cm.
• File origin: Born digital
• Access conditions: Open access
• Created on: 2023-09-19 13:49:01