Predicting programming assignment difficulty

Show full item record



Permalink

http://urn.fi/URN:NBN:fi:hulib-201908133223
Title: Predicting programming assignment difficulty
Author: Harju, Esa
Contributor: University of Helsinki, Faculty of Science
Publisher: Helsingin yliopisto
Date: 2019
Language: eng
URI: http://urn.fi/URN:NBN:fi:hulib-201908133223
http://hdl.handle.net/10138/304687
Thesis level: master's thesis
Degree program: Datatieteen maisteriohjelma
Master's Programme in Data Science
Magisterprogrammet i data science
Specialisation: ei opintosuuntaa
no specialization
ingen studieinriktning
Discipline: none
Abstract: Teaching programming is increasingly more widespread and starts at primary school level on some countries. Part of that teaching consist of students writing small programs that will demonstrate learned theory and how various things fit together to form a functional program. Multiple studies indicate that programming is difficult skill to learn and master. Some part of difficulty comes from plethora of concepts that students are expected to learn in relatively short time. Part of practicing to write programs involves feedback, which aids students’ learning of assignment’s topic, and motivation, which encourages students to continue the course and their studies. For feedback it would be helpful to know students’ opinion of a programming assignment difficulty. There are few studies that attempt to find out if there is correlation between metrics that are obtained from students’ writing a program and their reported difficulty of it. These studies use statistical models on data after the course is over. This leads to an idea if such a thing could be done while students are working on programming assignments. To do this some sort of machine learning model would be possible solution but as of now no such models exist. Due to this we will utilize idea from one of these studies to create a model, which could do such prediction. We then improve that model, which is coarse, with two additional models that are more tailored for the job. Our main results indicate that this kind of models show promise in their prediction of a programming assignment difficulty based on collected metrics. With further work these models could provide indication of a student struggling on some assignment. Using this kind of model as part of existing tools we could provide a student subtle help before his frustration grows too much. Further down the road such a model could be used to provide further exercises, if need by a student, or progress forward once he masters certain topic.


Files in this item

Total number of downloads: Loading...

Files Size Format View
Harju_Esa_pro_gradu_2019.pdf 1.537Mb PDF View/Open

This item appears in the following Collection(s)

Show full item record