Enhancing Bayesian Neural Networks with Synthetic and Real Data

Edris Loftpouri MFE /24 discusses his interest on the implementation of Bayesian Neural Networks (BNNs) for macroeconomic forecasting. He also touches on Castastrophe Modeling https://www.linkedin.com/in/patrick-z-08bb5b5a/ https://www.linkedin.com/company/lehigh-master-in-financial-engineering/ This project develops a Bayesian Neural Network (BNN) for macroeconomic forecasting, using stochastic volatility and Bayesian shrinkage priors to manage complex, high-dimensional data. With layer-specific and neuron-specific activation functions, the model captures both long-term dependencies and short-term nonlinear dynamics. Offering adaptive uncertainty quantification and robust volatility handling, it’s ideal for risk analysis, economic policy, and quantitative finance applications. https://www.linkedin.com/in/edris-lotfpouri/