Dissertation defence (Computer Science): FM Erno Lokkila
Time
16.2.2023 at 12.00 - 16.00
FM Erno Lokkila defends his dissertation in Computer Science entitled “Understanding novice programmer behavior on introductory courses - Learning analytics approach” at the University of Turku on 16 February 2023 at 12pm (University of Turku, Main Building, Tauno Nurmi lecture hall, Turku).
Opponent: Assistant Professor Brett Becker (University College Dublin, Ireland)
Custos: Adjunct Professor Mikko-Jussi Laakso (University of Turku)
Digital copy of the thesis at UTUPub: https://urn.fi/URN:ISBN:978-951-29-9148-8 (copy the link to the browser).
***
Summary of the Doctoral Dissertation:
Technology offers many new possibilities for teaching. Learning analytics is one of these, and it focuses on analyzing, interpreting and utilizing the data generated by learners in improving teaching. In this thesis, I present a model which can be used to automatically identify novice programmer behavior.
The model consists of a state machine, which consists the possible states and state transitions possible by a student when doing an assignment. The student traverses through these states and all transitions are stored. Once the student finishes, the state transitions can be used to generate a Markov Model, which allows the mathematical modeling of the student. In the thesis, this model is used with machine learning algorithms to successfully cluster students into three distinct groups based on their behavior.
The data generated by the model can also be used to estimate the current skill level of the student programmer. The student’s learning can be tracked and with the data from the model, teachers can offer targeted help to students who need it, even on courses with hundreds or thousands of students. This research also paves the way for creating assignments that automatically adapt to the current skill level of the student.
Opponent: Assistant Professor Brett Becker (University College Dublin, Ireland)
Custos: Adjunct Professor Mikko-Jussi Laakso (University of Turku)
Digital copy of the thesis at UTUPub: https://urn.fi/URN:ISBN:978-951-29-9148-8 (copy the link to the browser).
***
Summary of the Doctoral Dissertation:
Technology offers many new possibilities for teaching. Learning analytics is one of these, and it focuses on analyzing, interpreting and utilizing the data generated by learners in improving teaching. In this thesis, I present a model which can be used to automatically identify novice programmer behavior.
The model consists of a state machine, which consists the possible states and state transitions possible by a student when doing an assignment. The student traverses through these states and all transitions are stored. Once the student finishes, the state transitions can be used to generate a Markov Model, which allows the mathematical modeling of the student. In the thesis, this model is used with machine learning algorithms to successfully cluster students into three distinct groups based on their behavior.
The data generated by the model can also be used to estimate the current skill level of the student programmer. The student’s learning can be tracked and with the data from the model, teachers can offer targeted help to students who need it, even on courses with hundreds or thousands of students. This research also paves the way for creating assignments that automatically adapt to the current skill level of the student.
Additional information
University Communications