PhD Candidate in Methodology and Statistics
Leiden University, Leiden
I develop methods that help behavioral researchers interpret machine learning findings and combine evidence across studies. My PhD research builds on RuleSHAP, which combines interpretable prediction rules, Shapley values, and Bayesian uncertainty estimates. I aim to extend this framework to multilevel data and derive standardized effect sizes for meta-analysis, so that complex patterns discovered by machine learning can contribute to cumulative knowledge about human behavior.
Course Project · Radboud University · 2026
Analyzed prediction market accuracy using 116,000+ resolved Polymarket markets from 2025. Evaluated forecast calibration across seven domains (Sports, Crypto, Politics, etc.) via Brier Score decomposition.
Master's Thesis · Radboud University · 2025 – 2026
Challenging the assumption of risk-taking as a static trait. Developed a computational pipeline using Auxiliary Particle Filters to recover individual trial-by-trial trajectories of latent risk sensitivity.
Research Project Co-Lead · Radboud University · 2024 – 2025
Engineered a continuous joystick-based Approach-Avoidance Task in PsychoPy with millisecond precision to investigate cognitive mechanisms underlying academic anxiety.
Position Paper (Course Project) · Radboud University · 2024
Proposed a predictive coding account to reconcile conflicting evidence on whether naming emotions dampens or intensifies them. Argued that prediction uncertainty is the key mechanism, modulated by label-experience intensity mismatch and labeling freedom.
Research Project Lead · Sun Yat-sen University · 2022 – 2024
Applied NLP to extract sentiment from 53,000+ crowdfunding texts. Constructed a novel Emotional Co-expression Network framework to quantify structural properties of emotional information.