Proyecto I+D
Project PID2022-136252NB-I00 MIDAMALERECO
SPATIOTEMPORAL MICRODATA. MACHINE LEARNING ALGORITHMS AND COMPLEX NETWORKS IN SOCIOECONOMIC
With a Public character. It has been granted under a regime of Competitive.
The project outlined in this proposal is centered on analyzing spatio-temporal data, georeferenced at the point level, from methodological and applied perspectives. We will make use of cutting-edge tools: recursive partitioning and Machine Learning algorithms and measures coming from information theory based on entropy, all of which will be used in conjunction with the classic tools of spatial data analysis. The project will continue with the highly productive scientific work of this research group, which has received continuous funding from this call for the last 15 years (PID2019-107800GB-I00; ECO2015-65758-P; ECO2012-36032-03: EC02009-10534/ECON). The proposal addresses four general objectives. The first objective is to combine Machine Learning algorithms with classic spatial econometric methodologies to improve modeling. This entails increasing the models predictive power and aiding in selecting the most adequate spatial econometric model. The second objective is focused on studying spatio-temporal processes with network and panel structures, and developing methodologies to determine: (i) time events not observable from a a large panel; (ii) causal relationships among the nodes of a complex grid where the effects of confounding factors and unobserved latent variables are mitigated, and (iii) the design of a universal model of migration flows brought about by climate change. The third objective is to measure the impact of environmental deterioration on the relat
Researchers
