Min-seok Kim

I am a graduate student in the Department of Applied Statistics at Chung-Ang University, advised by Prof. Jooyoung Lee. My research focuses on bridging the gap between Rigorous Statistical Theory and Advanced Machine Learning to solve complex, high-dimensional data challenges.

I am particularly passionate about modeling Structured Data—ranging from spatial trajectories and functional signals to complex relational graphs. My current work involves:

🌐 Graph Neural Networks (GNNs) — Developing Graph Attention Network (GAT)-based survival models. I integrate deep representation learning with Cox regression to capture non-linear interactions within networked data, such as P2P lending default risks.

📈 Functional & Spatial Data Analysis — Applying the Square-Root Velocity Framework (SRVF) and MFPCA to analyze time-varying trajectories. I have experience handling spatial coordinate systems (WGS84/EPSG:5179) and high-frequency sensor data, ensuring data alignment and registration for robust feature extraction.

⚖️ Causal Inference & Clinical Research — Implementing Target Trial Emulation to mitigate biases (e.g., immortal time bias) in real-world evidence. I focus on making ML models not just predictive, but interpretably causal for decision-making in healthcare and policy.

With a solid foundation in Applied Statistics (B.S.), I leverage Python, R, and SQL (handling large-scale datasets like NHIS) to transform raw, noisy data into actionable insights. I am always looking for ways to apply these methodologies to new frontiers like Mobility (SDV) and Quantitative Finance.