Research

My research program is organized around four pillars that share a common methodological core: modeling complex temporal and spatial patterns in data to produce actionable insights.

Financial Volatility & Time Series Econometrics

This is my primary research area. I develop new econometric models for financial volatility that capture how markets encode, retain, and forget information over time.

The Shape of Volatility Memory — My central project establishes that volatility memory kernels across 100+ financial assets (equities, currencies, commodities, fixed income, crypto, VIX) follow sub-exponential decay patterns better described by stretched exponentials than by the standard GARCH or FIGARCH alternatives. This work introduces the GMARCH and SEARCH model frameworks. In production — target: Journal of Business & Economic Statistics

MF2V-GARCH — Augments Conrad & Engle’s (2025) Multiplicative Factor Multi-Frequency GARCH with a smoothed trading volume component, estimated across 15 US assets with rolling out-of-sample evaluation. Includes a complete MATLAB toolbox. In production — target: Journal of Forecasting

MF2-GARCH-A — Extends the MF2-GARCH with a sign-sensitive long-run component that captures asymmetric responses to positive and negative return shocks. Near completion — target: International Journal of Forecasting

MF2-EGARCH — An exponential GARCH version of the multiplicative factor multi-frequency framework, with several variations for the long-term component including an asymmetric MEM and a logarithmic MEM. Near completion — target: International Journal of Forecasting

Additional models: Multiple Regime Hyperbolic GARCH (MR-HYGARCH), Fractionally Integrated GJR-GARCH (FI-GJRGARCH), and Smooth Transition FI-GJRGARCH extensions.

Software

Hybrid Time Series & Machine Learning

I am exploring the intersection of classical time series econometrics and modern deep learning, with the goal of building hybrid models that combine the interpretability of statistical methods with the flexibility of neural architectures. This work aims to produce explainable ML-based forecasting models grounded in rigorous statistical foundations — bridging the gap between traditional econometric approaches and data-driven methods.

GRU-CNN Hybrid for Influenza Forecasting — A hybrid gated recurrent unit / convolutional neural network model with Yeo-Johnson scaling for forecasting influenza-like illnesses. Presented at JSM 2025 (Infectious Disease Epidemiology Section). Joint work with student Noah Gallego and Prof. Isuru Ratnayake (KUMC). Manuscript in preparation

Environmental Justice & Air Quality

I pursue data-driven applications that serve public needs, with current focus on air pollution exposure disparities.

Income-Based Exposure Disparities in California — Introduces a reproducible, county-level, multi-pollutant statistical framework (PM2.5, NO₂, O₃, SO₂, CO) to quantify long-term income-based exposure disparities across California counties. Joint work with student co-authors Kayla Ko and Christian Rodriguez. Funded by the CES Mini-Grant. Published in Environmental Research: Health, 2025

Fine-Scale Air-Quality Heterogeneity in Twin Cities — Analysis of data from a dense low-cost network of 45 multi-pollutant sensors in the Minneapolis-St. Paul metropolitan area to investigate spatiotemporal pollutant patterns and exposure disparities. Published in ACS ES&T Air, 2025

Air Pollution & Fertility Patterns — Ongoing interdisciplinary project investigating relationships between pollution and fertility in California counties, combining statistical modeling and machine learning. CV Pathway 2025 Summer Research Apprenticeship.

AI in Statistics Education

I maintain a growing research agenda at the intersection of artificial intelligence and statistics education, supported by the California Learning Lab ELEVATE grant.

Scaffolded LLM Tutors — Lead author on “From Answer Generators to Thinking Partners,” which develops a framework for AI-driven tutoring, prompt engineering, and ethical classroom integration using Custom GPTs. Submitted to JSDSE

Structured AI Tutoring in Engineering Education — Co-authored WIP paper contributing AI prompt design methodology. Published — IEEE Frontiers in Education 2025

Computer-Aided Instruction for K–12 Teachers — Cognitive apprenticeship approach to LLM integration for K–12 professional development. Published — IEEE FIE 2025

Collaborators

  • V.A. Samaranayake — Missouri University of Science and Technology (Ph.D. advisor)
  • Isuru Ratnayake — University of Kansas Medical Center
  • Alberto Cruz — CSUB, Computer & Electrical Engineering (ELEVATE PI)
  • Jianjun Wang — CSUB, Mathematics
  • Maruti Mishra — CSUB
  • Eduardo Montoya — CSUB, Mathematics (Statistical Data Science curriculum)
  • Bilin Zeng — CSUB, Mathematics

Grants & Funding

Year Grant Role Source
2025 Air Pollution Disparities in California, 2000–2025 PI CES Mini-Grant, CSUB
2024–25 ELEVATE: Enhancing Learning Experiences Via AI Techniques Co-PI California Learning Lab AI FAST Challenge