SUMMARY blog!!

It is such a pleasure for me to get involved into the summer research at W&M. This is also my first formal systematic and academic research experience at college. Thank to professor Frazier, I learned how to conduct an academical research professionally from the very beginning, searching for an area, and on the way to the end, making summary and highlight the crucial and creative points in the research. This is a great treasure for me. This research is mostly about using random forest and mathematical models including hierarchal Bayesian model to generate a synthetic population description for Liberia, and the future work would be applying the method into other regions and countries and also to solve problems related to various fields, like public health and disease tracking.

Here is a brief summary of the research. Since global patterns of population change highlight the need for spatially explicit and comparable high-resolution gridded population datasets that accurately depict the spatial distribution of the residential human population and inform many fields, including infectious disease assessment, disaster response, adaptive strategies towards climate change migration, and many of the Millennium Development Goals. These kind of research is critical in many health, economic, and environmental fields across various temporal and spatial scales. (High-resolution gridded population datasets for Latin America and the Caribbean in 2010, 2015, and 2020). The methods related are mostly about Random forest – based asymmetric population mapping approach, random forest simulation, and bottom-up and top-down approaches. We compared these methods and random forest is a new and improved one from the original ones. For mathematical models, since HBM and multinomial logistic regression model are useful in converting the discrete data information into more continuous and detailed one, they are studied more during the summer. We have to use the centre points from the map generated by random forest to map the original sample data from DHS and distribute the data within the region around centre points and then use HBM to generalise the data into more detailed regions. Also, we might use other methods to make dynamic maps to track population migration and to predict the spread of disease.

Again, I am so happy to get involved into this great research and look forward to do more about it. Also, I hope this can be applied into broad range of fields in the near future!