Quantile Regression
High-dimensional Statistics
Financial Machine Learning
Time Series Analysis
Machine learning with Microstructure measures on High-frequency Trade data
Collaboration with Professor David Easley and Professor Sumanta Basu
Mine recent high-frequency trade data
Create microstructure measures with High-frequency Trade data
Deal various questions on prediction accuracy of different time bars and different machine learning algorithms
A Pathwise Coordinate Descent algorithm for High-dimensional Quantile Regression
Create a new, exact, simple, and fast coordinate descent algorithm for high-dimensional data using Gauss-Seidel method
Experiment different penalty functions, including non-convex SCAD and MCP
Learning Financial Networks with High-Frequency Trade Data
Data Science in Science, https://doi.org/10.1080/26941899.2023.2166624
Utilize High frequency trade data
Compare the robustness of Random Forest compared to Gradient Boosting and Deep Neural Networks
Return prediction by machine learning for the Korean stock market
Journal of the Korean Statistical Society, https://doi.org/10.1007/s42952-023-00245-0
Utilize the research by Gu et al. (2020) to verify the machine learning method which is capable of asset pricing prediction
Conducted analysis methods including OLS, PCR, PLS, Random Forest, Gradient Boost, and Neural Network on American CRSP data (all characteristic variables and macroeconomic variables) utilized on the researches
Worked on setting the Korean stock data as identical as possible and applied above methodologies
In charge of Python-based data preprocessing
Cyber risk management via Loss Distribution Approach and GARCH model combined with Extreme Value Theory
CSAM(Communications for Statistical Applications and Methods), https://doi.org/10.29220/CSAM.2023.30.1.075
Use the loss distribution approach (LDA) and the time series model to describe cyber losses of financial and non-financial business sectors, provided in SASⓇ OpRisk Global Data
Measure value at risk (VaR) by combining Peaks over threshold (POT) method
Build a two-dimensional model reflecting the dependence structure between financial and non-financial sectors through a bivariate copula and checked the model adequacy through VaR