with Svetlana Ryabova and Viktor Glebov
A substantial amount of data on COVID-19 has been collected in the US and around the world. Our inter-disciplinary team of computer scientists, an applied cultural anthropologist, and a chemist will use machine learning and artificial intelligence techniques9 to distill salient features from existing datasets. After development of models based on large populations, we plan to test them on small isolated populations of different ethnicity and socioeconomic background. Specifically, Navajo residing on the reservation.
The Navajo nation made the national news due to its extraordinary high rate of COVID-19 infection and death. Back in April and May only New Jersey and New York have higher per-capita COVID-19 infections than the Navajo Nation. Navajos are at higher risk from COVID-19 because the disease factors include geographical isolation, limited access to health care facilities, elevated poverty levels, and high levels of diabetes and other pre-existing conditions. In addition, decades of uranium and other extraction activities have caused many Navajos to suffer from major autoimmune disorders that in turn have left many of them vulnerable to the COVID-19. Thus, the problem of COVID-19 spread in the Navajo reservation is of severe urgency and requires a rapid response.
Meantime, as the COVID-19 pandemic emerged and spread in the US and around the world, scientists, employing clinical and non-clinical research approaches, started working on possible ways to explain and reduce the outbreak, employing clinical and non-clinical research. With the help of NFS RAPID grant (NSF DMR-2031548) that is under an existing NSF PREM grant (NSF DMR-1523611) our New Mexico Highlands University interdisciplinary team of faculty and graduate students from computer science, anthropology, and chemistry set out to collect publicly available data and use machine learning methods to explain existing differences in the spread of COVID-19 within the United States in general, and develop a model that assesses and predicts current and future outbreaks of the disease on the Navajo reservation. The team collects data and tests various mathematical and computer models on data that ranges from a variety of COVID-19 epidemic models from the United States and around the world, through socioeconomic and demographic characteristics, population density, and so on. The team will ultimately incorporate data that is specific to the Navajo Nation and unique to Navajo cultural and socio-economic characteristics (multi-family households, lack of running water and/or electricity, distance from health care facilities).
Our inter-disciplinary team will use comprehensive data-demanding models and artificial intelligence methods, such as artificial neural networks to study the spread of COVID-19 on the Navajo reservation. Ultimately, the outcome will help determine and predict disease spread on the Navajo reservation. It will, in turn, allow for the optimal timing of efficient anti-epidemic measures that will simultaneously allow for preventing and minimizing the spread of infection hospitalization and death, and associated economic impacts on the Navajo Nation.
Utilizing a diverse set of methods found in machine learning will enable our team to find very robust models that have both predictive and classification capabilities across many domains. We are working on creating datasets that focus on the Navajo Nation that will include socio-cultural features of these communities and compare results on virus spread depending on different introductory factors. We hope that such approach will help to evaluate situation in these areas and help to produce specific recommendations. We plan to turn over our findings to the Navajo Nation. For additional information please refer to https://www.nmhu.edu/hu-computer-science-grant-to-help-navajo-nation-with-covid-19-data/.