Software developer and researcher

Research focus

How can complementary detection methods help identify suspicious patterns while keeping model decisions and evaluation trade-offs interpretable?

Methods

A rule-based baseline, Isolation Forest, and logistic regression, alongside Bayesian uncertainty, bootstrap analysis, and Monte Carlo methods.

Evaluation & limitations

The evaluation compares detection performance with the rule-only baseline. Results depend on the dataset and sampling assumptions; experimental performance does not establish readiness for deployment in financial systems.

Research focus

How does the fidelity of a flight-booking prototype affect the experience of using it?

Study design

A study with 25 participants comparing three prototype formats: paper sketches, wireframes, and an HTML/CSS interface.

Evaluation

Measures included task completion time, errors, NASA-TLX cognitive workload, and mental-model alignment. The study considers both task performance and participants’ experience of the interface.

Publication

Evaluating Prototype Fidelity: Impacts on Cognitive Workload and Mental-Model Alignment in Flight-Booking Interfaces. International Journal of Research in Interdisciplinary Studies, 3(11), 9–14, November 2025.

Research focus

How do different approaches to emotional adaptivity shape the experience of using a study tool?

Study design

A mixed-methods, within-subjects study with 17 participants, comparing self-report, automatic emotion detection, and standard timer conditions using objective and subjective measures.

Findings

Participants preferred the self-report mode for its sense of autonomy, control, and accuracy. Automatic detection was considered helpful but raised trust concerns, highlighting the importance of agency and transparency in affective interfaces.

Publication

Understanding Emotional Adaptivity in Study Tools: A Mixed-Methods Evaluation of Eunoia. International Journal of Research in Interdisciplinary Studies, 3(11), 27–35, November 2025.