Evolving Spatially Aggregated Features from Satellite Imagery for Regional Modeling
This work addresses the need for actionable regional insights from satellite data for domain-specific applications like snow property prediction, representing an incremental advancement in feature construction methods.
The paper tackled the problem of deriving regional summaries from high-resolution satellite imagery for geospatial modeling by proposing a method to induce spatial aggregations as part of the machine learning process, resulting in features driven by prediction performance rather than prior assumptions, with Genetic Programming showing effectiveness in synthesizing aggregations and improving predictive models in experiments on snow properties in high-mountain Asia.
Satellite imagery and remote sensing provide explanatory variables at relatively high resolutions for modeling geospatial phenomena, yet regional summaries are often desirable for analysis and actionable insight. In this paper, we propose a novel method of inducing spatial aggregations as a component of the machine learning process, yielding regional model features whose construction is driven by model prediction performance rather than prior assumptions. Our results demonstrate that Genetic Programming is particularly well suited to this type of feature construction because it can automatically synthesize appropriate aggregations, as well as better incorporate them into predictive models compared to other regression methods we tested. In our experiments we consider a specific problem instance and real-world dataset relevant to predicting snow properties in high-mountain Asia.