Back

Prediction of Covid-19 Infections Through December 2020 for 10 US States Incorporating Outdoor Temperature and School Re-Opening Effects-August Update

Newell, T. A.

2020-09-18 epidemiology
10.1101/2020.09.14.20193821 medRxiv
Show abstract

End-of-August updates for Covid-19 infection case predictions for 10 US States (NY, WA, GA, IL, MN, FL, OH, MI, CA, and NC) are compared to actual data. Several states that experienced significant summer surges gained control of accelerating infection spread during August. The US as a whole and the 10 States investigated continue to follow periods of linear infection growth that defines a boundary separating accelerated infection growth and infection decay. August 31 predictions (initiated July 27, 2020) for four States (NY, WA, MI and MN) are within 10% of actual data. Predictions for four other States (GA, IL, CA, and OH) agree between 10 and 20% of actual data. Predictions for two States (FL and NC) are greater than 20% different from actual data. Systematic differences between predictions and actual data are related to the impact of the June-July summer surge, and human behavior reactions (ie, increased face mask usage and distancing) to accelerated infection growth. Outdoor temperature effects and school re-opening effects are not apparent nor expected for August. Human behavior parameters (Social Distance Index, SDI, and disease transmission efficiency, G, parameters) are adjusted to mirror August data. Comparisons of actual versus predicted daily new infection cases display the complexity of SARS-CoV-2 transmission.

Matching journals

The top 7 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.