## Tokyo: New positive cases

(Update only the reported number of new positive cases)

Optimistic scenario: Vaccination steadily increases from now on (at a pace of a million shots per day as of February 22). Flows of people do not come back.

Pessimistic scenario: Vaccination level stays the same. Flows of people come back in March (the data is based on that of last year)

Input actual data as-needed. The data of 2/14 does not include the effect of three consecutive holidays because it is the data until February 10.

In Tokyo, the flows of people on February 10 (Thursday) drops by more than 5%. Need to focus on whether the number of new positive cases will drop due to factors such as the weather (snow) (Reference: Dr. Ohsawa’s discussion on the fifth wave attenuation).

Input data of March is set by referring to 2021. It remains high partly due to the consideration of the trend of the flow and behavior of people coming back.

## Effect if quasi-state of emergency is lifted on 3/6

*Because it is based on the past data, the effect of transition to having a meal without a mask is not included.

## Inflow of BA2 subvariant

Based on the reported value, BA2 is assumed to be 1.3% and 4.2% of infected people as of 2/14 and 2/21, respectively (approximated from the significant digit of the reported number of Tokyo). Predicted based on the optimistic scenario on page 2.

The assumption is 1.5 times stronger infectability and the same generation time as BA1

https://www.bousai.metro.tokyo.lg.jp/_res/projects/default_project/_page_/001/021/087/80/20220225_10.pdf

*Projection example based on the assumption (lower accuracy than other predictions).

If the generation time of BA2 is shorter, it will be smaller

## Tokyo: Number of newly hospitalized patients, number of seriously ill patients

As for the number of newly hospitalized patients,

there is almost no difference from that of 2/14

*The definition of the number of seriously ill patients is that of the Tokyo Metropolitan Government (https://stopcovid19.metro.tokyo.lg.jp/monitoring)

## Tokyo: Number of fatalities Update

*Accuracy is lower compared with the number of new positive cases, number of new positive cases and number of seriously ill patients (the residual is larger in terms of value).

Estimation may not be able to fully reflect the deterioration in a patient’s condition such as complications, as well as the number of new positive cases.

**Need to closely monitor the situation, as the predicted values of the number of seriously ill patients and number of fatalities might depend on the effectiveness of vaccination.

The effect of quasi-state of emergency is small (until the end of March).

## Reposted: Prediction system based on AI (deep learning)

Based on deep learning (LSTM model), directly predicted the number of new positive cases/number of seriously ill patients per day (one-week average value), etc. To identify the input parameters, things other than the future number of positive cases should be known.

The data volume in appearance can be increased by using standardized data instead of that of prefectures. Estimate based on non-linear regression (the number of parameters is virtually unlimited)

*The values can be calculated within several months to the extent that the accuracy can be secured

1.E. A. Rashed and A. Hirata, “Infectivity upsurge by COVID-19 viral variants in Japan: evidence from a deep learning modeling.” Int. J. Environ. Res. Public Health, 2021.

## Reposted: Estimate example of the 4th and 5th waves in Tokyo and Osaka

(Learning up to April 15, 2021 for Tokyo and March 4, 2021 for Osaka, and estimate for the following period)

Learned with the estimate period and verified the effectiveness of the constructed network with input data of things other than the number of new positive cases being known

Right axis: Standardized value of flows of people at stations, infectability of the virus, and Twitter (drinking party) data

*Prepared using the Twitter data provided from NTT DATA by Toyoda Lab, Institute of Industrial Science, The University of Tokyo

## Reposted: Estimate example of the 4th and 5th waves in Tokyo

(Learning up to April 15, 2021 for Tokyo)

Learned with the estimate period with input data of things other than the number of new positive case being known. Searched input parameters with strong relationships with the number of new positive cases

The nighttime retention population is not necessarily an optimal indicator for its estimation even if it has a causal relationship with the number of new positive cases.

*The other input parameters are being verified

Right axis: Standardized value of flows of people at stations, infectability of the virus, and Twitter (drinking party) data