Research

Atmosphere & Environment Sciences

Title :

Hybrid-AI Quantile Regression Combining Nonparametric Statistical Methods and Physics-Informed Neural Networks for Analysing Spatio-temporal Climate Data

Area of research :

Atmosphere & Environment Sciences, Earth

Focus area :

Climate Change, Digital technologies

Principal Investigator :

Prof. Soudeep Deb, Indian Institute Of Management, Bangalore, Karnataka (560076)

Timeline Start Year :

2024

Timeline End Year :

2026

Contact info :

Details

Executive Summary :

Climate change and rapid urbanization pose significant threats to urban environments, leading to unpredictable patterns and extreme events. Researchers aim to develop advanced artificial intelligence AI models for spatio-temporal data analysis and prediction to contribute to the United Nations Sustainable Development Goal 11 Sustainable Cities and Communities. The project aims to develop a versatile and generic approach for estimating and predicting spatio-temporal quantiles in climate data, combining nonparametric methods with physics-informed machine learning techniques. The methodology will quantify and account for uncertainties in data and models, providing reliable prediction intervals for estimated quantiles. The methodology will demonstrate applicability to real-world data, aiding policymakers and urban planners in making data-driven decisions for sustainable development and climate change adaptation. The research will begin with preprocessing and analyzing historical spatio-temporal rainfall data, focusing on nonparametric quantile regression techniques and deep learning-based methods. The proposed method will be adapted and applicable across various domains, enabling stakeholders to monitor spatio-temporal patterns and explore what-if scenarios to mitigate climate change and urbanization impacts.

Total Budget (INR):

22,38,766

Organizations involved