I work at the intersection of financial econometrics, high-frequency financial time series, and machine learning, with a focus on robust and leakage-free empirical modeling.
My current research explores how deep sequence architectures can model noisy, non-stationary financial time series and how model performance changes across different volatility and market regimes.
- Financial Machine Learning
- Financial Econometrics
- High-Frequency Financial Time Series
- Deep Sequence Modeling
- Volatility and Regime Modeling
- Risk-Aware Forecasting
- Leakage-Free / Causal Evaluation
- Machine Learning for Financial Markets
I study 5-minute XAUUSD financial time series using chronological and walk-forward evaluation frameworks.
Current work includes the empirical comparison of:
- TimesNet
- PatchTST
- LSTM / GRU
- TCN
- XGBoost
- Random Forest
The research focuses on temporal dependencies, volatility regimes, risk-aware labeling, and realistic out-of-sample evaluation.
Deep Sequence Architectures in High-Frequency Finance: Benchmarking TimesNet and Classical Machine Learning Across Volatility Regimes
SSRN:
https://papers.ssrn.com/sol3/papers.cfm?abstract_id=7034100
Financial Econometrics
- Time-Series Analysis
- Volatility Modeling
- Regime Detection
- Risk Modeling
- Walk-Forward Evaluation
Machine Learning
- PyTorch
- Scikit-learn
- XGBoost
- Classical ML
- Deep Sequence Models
- Transformers
- LSTM / GRU
- TimesNet
- PatchTST
Computational Research
- Python
- NumPy
- Pandas
- CuPy
- GPU-accelerated computation
- Reproducible experimentation
- Git / GitHub
Research implementation of deep sequence models for high-frequency financial time-series analysis.
Classical machine-learning benchmarks and evaluation pipelines for financial time-series research.
Computational framework for scalable feature engineering and experimentation on large financial datasets.
M.Sc. Business Administration – Financial Management
Yazd University, Iran
Graduate-Level Data Science Coursework — 87 ECTS
University of Verona, Italy
The Data Science programme was taught entirely in English.
🔗 LinkedIn: Mohammadreza Akhlaghi