## Prediction model

１．Predict the number of new positive cases from the Japanese situation considering the impact of the variant (no definition of the vaccination rate)

２．Learned and corrected the data of other countries to consider the impact of vaccination in more detail

Model (1): The impact of vaccination is not defined explicitly

- The input values include the weather information (minimum/maximum and average temperature), data on flows of people, weekdays or holidays, presence or non-presence of declaration of a state of emergency, and number of new positive cases to date as well as the newly added variation label (0: Standard, 1: alpha, 2: delta)

- The future values of weather data and the data of flows of people are assumed to be similar to those of last year

- The applicable city of learning data is Tokyo

Model (2)

- Separated the prediction model to make it possible to consider only the impact of vaccination

- The input values of the network are the number of new positive cases and the effectiveness of vaccination

- The applicable cities of learning data are Tokyo and Tel Aviv

- The age composition is not considered (no data of Tel Aviv)

## Model of the infection-prevention effect of vaccination

Based on the medium model (infection-prevention effects of the 1st and 2nd vaccinations: 65% and 75%, respectively) and the report of 10/5, it is assumed that the prevention effect of the 3rd vaccination is 95% and the attenuation rate of the prevention effects of the 1st and 2nd vaccinations of half a year later are 35% and 55% (decrease linearly). The vaccination effect is assumed to keep its peak for 14 days after vaccination and then attenuate (Reference:https://doi.org/10.1038/d41586-021-02261-8）。

## Twenty-seven patterns of vaccination scenarios

Vaccination scenarios considered this time

- Second vaccination rate up to the end of the year: 75%, 80%, and 85%

- Start timing of the third vaccination: Early December (Dec 1), early January, and early February

- Daily number of people vaccinated three times: 80,000, which is the baseline (equivalent to 7/1), 1/2 of the baseline (40,000), and 1/3 (25,000)

Based on the medium model (infection-prevention effects of the 1st and 2nd vaccinations: 65% and 75%, respectively) and the report of 10/5, it is assumed that the prevention effect of the 3rd vaccination is 95% and the attenuation rate of the prevention effects of the 1st and 2nd vaccinations of half a year later are 35% and 55% (decrease linearly). The vaccination effect is assumed to keep its peak for 14 days after vaccination and then attenuate (Reference:https://doi.org/10.1038/d41586-021-02261-8）。

## List of prediction results in 27 patterns of vaccination scenarios

Number of new positive cases

## If the second vaccination rate at the end of the year is 75%

*For reference, the case without the third vaccination was also considered

## If the second vaccination rate at the end of the year is 80%

*For reference, the case without the third vaccination was also considered

## If the second vaccination rate at the end of the year is 85%

*For reference, the case without the third vaccination was also considered

## Prediction of the number of seriously ill patients in Tokyo

Sc9 as the worst case (considered assuming that the second vaccination rate at the end of the year would be 75%, the third vaccination would be started on February 1, and the vaccination speed would be 1/3 (25,000 people/day))

Even without considering the third vaccination, the above result was almost not affected

and an increase in the number of seriously ill patients was not predicted.

## Considering the impact of flows of people (based on the current flows of people in Tokyo)

The worst scenarios are those with three different vaccination speeds assuming that the second vaccination rate would be 80% and the third vaccination would start on December 1 (Scenarios 10, 11, and 12)

## If flows of people recover the level equivalent to that before the coronavirus catastrophe

If all flows of people are equivalent to the base value (five weeks between January 3 and February 6, 2020, median of each day of the week)

## Impact of introducing the vaccine passport

Based on the effective reproduction number (R) in London reported on September 27, the effective reproduction number is assumed to rise by 10–30% because of the introduction of the vaccine passport if flows of people do not change. There was almost no change in Singapore.

This method is based on learning of actual data, so detailed parameters cannot be set.

## Summary

- Machine-learned the relationship between the number of new positive cases and the vaccination rate in Tokyo and applied it to the estimate of the number of new positive cases in Israel (Tokyo -> Israel).

- Estimated the new infection-prevention effect of vaccination and its temporal transition.

- Approximated the attenuation of the effect of the second vaccination and the effect of the third one based on the medium model (first: 65%, second: 75%).

- Estimated the effect of the second vaccination (75%) to decrease to 35% in half a year and the infection-prevention effect of the third one to be 95%.

The increase in the number of new positive cases after the year-end and new-year holidays is dominant (learning of the case with small flows of people).

January 2022 and onward: Even if it increases, it will soon start to decrease thanks to the third vaccination. If it is not conducted, the transition will become flat and the decrease trend will be difficult to confirm.

The number of seriously ill patients is 30 or fewer.

Even if the third vaccination is conducted, the number of new positive cases will increase with complete lifting of restrictions.