Project Title

A Comparative Study of Economic Cost Prediction Accuracy for Tornado Damage: Classical vs. Hybrid (Classical+Quantum) Neural Networks.

Research Team

Contact Information

Phys. Jorge O. Cedeño & Fin. Omar Alsaid

Institutional e-mail:

Overview

Tornadoes are among the most frequent and devastating natural disasters worldwide. Although advances in weather forecasting have allowed the development of increasingly accurate technologies to anticipate these events, significant difficulties persist in the proper estimation of the associated economic damage, which complicates the planning and distribution of resources in emergency situations. In this research project, a hybrid neural network is developed to predict the economic cost of damage caused by tornadoes. For this purpose, a multi-output neural network was implemented using TensorFlow and Keras, capable of estimating economic losses from specific characteristics of tornadoes. Additionally, a quantum layer was incorporated using the PennyLane tool, with the aim of evaluating improvements in the model's accuracy. Preliminary results show a better performance of the hybrid (classical-quantum) model compared to traditional architectures, evidencing the potential of quantum computing as a complementary tool in complex climate prediction and impact assessment problems.

jorge.cedeno@czetesis.com

omar.alsaid@czetesis.com

Alternative e-mail:

jorgebectruc18@hotmail.com

omar-alsaid-sulaiman@outlook.com