:strip_icc()/i.s3.glbimg.com/v1/AUTH_59edd422c0c84a879bd37670ae4f538a/internal_photos/bs/2017/M/z/yB2xSbQpCJjkmg5SaNtQ/amis-ong-vila-andrade-credito-celso-tavares-g1-121-.jpg)
[ad_1]
Countries with great social inequality, such as Brazil, are among those with the highest mortality rate from Covid-19 and are also the ones that take the longest to see a reduction in contagion. This is what a study by researchers from the Federal University of Minas Gerais points out, of artificial intelligence tools.
Conducted in partnership with Kunumi, an organization dedicated to the development of emerging technologies, the survey indicates which variables led to a worsening or improvement of death rates in countries affected by the pandemic.
They were raised more than 200 indicators from 211 countries, extracted from official databases, such as Datasus, IBGE and others. To condense this data, an artificial intelligence solution was used, according to one of the researchers and head of the laboratory, Adriano Veloso.
“This model that we created is capable of combining all the variables, which is not trivial. We generate and train this model day after day, to anticipate or predict accelerations; when the outbreak would increase or decrease. This tool is still capable of automatically resetting itself in case of error, ”he said.
According to Adriano, the model anticipates the increase in the number of deaths and the main causes in each country in 15 days, who would be the “culprits” of this phenomenon.
“We create a model and anticipate in 15 days how many deaths will increase. And in doing so, the goal is to be able to explain why the acceleration will increase. The explanation comes in the form of “guilt”. In Brazil, the large number of people living in the favela and social inequality are the factors that lead to a large number of deaths, ”explained Adriano.
In European countries, for example, the large number of elderly population living in nursing homes would be the factor responsible for the number of deaths. “Belgium suffered a lot, if you consider that it does not have a very large population. Holland, Switzerland too. What our tool offered to explain is that the culprit is the number of nursing homes. These countries have many nursing homes. Nurses left the hospital, they stayed asymptomatic and they took him to the residents of the asylum, “he said.
July 1 – Elderly woman hugs the director of a Belgian nursing home – Photo: Yves Herman / Reuters
The proposal of the tool is, according to the researcher, Help public authorities identify what is driving the expansion or weakening of the pandemic., according to the situation of each location. “The tool aims to address the complexity to support public administrators of the problem and all the complexities,” said Gabriella Seiller, Kunumi executive.
The survey and data are available for viewing on the website.
Sars CoV-2: Viral load can predict mortality – Photo: Maurizio De Angelis / Science Photo Library
For the first time, in mid-August, the platform identified a slowdown in the contagion of Covid-19 in Brazil. However, the researcher sees the good news with reservations.
“What we’re seeing in the tool is that the acceleration is dropping below zero. The velocity of propagation is starting to slow down. This is a very good thing. However, when we see the importance of the variables that are slowing down the spread, which would be the closure of schools, we are concerned when we see that some states are resuming classes in the classroom. This relaxation, not only by the government, but also by the population that does not respect it, is a danger. ”
