Table 6 Top 5 countries (more than 10,000 total cases over prediction period) with best and worst performance by percentage error with Models 1 and 2.
From: A novel bidirectional LSTM deep learning approach for COVID-19 forecasting
Model 1 | |||||
|---|---|---|---|---|---|
Top 5 countries with best results | Average cases per day during prediction period | In or out of infection wave | Part of infection wave | Number of infection waves | |
1 | Montenegro | 438 | In | Increasing | 4 |
2 | Belgium | 2150 | Out | NA | 3 |
3 | Slovakia | 2120 | In | Declining | 2 |
4 | Paraguay | 836 | In | Increasing | 2 |
5 | Estonia | 507 | In | Declining | 2 |
Top 5 countries with worst results | Average cases per day during prediction period | In or out of infection wave | Part of infection wave | Number of Infection waves | |
|---|---|---|---|---|---|
1 | Turkey | 7386 | Out | NA | 1 |
2 | Denmark | 869 | In | Declining | 2 |
3 | Malawi | 697 | In | Increasing-peak-declining | 1 |
4 | Ghana | 502 | In | Increasing | 2 |
5 | Mozambique | 790 | In | Increasing | 2 |
Model 2 | |||||
|---|---|---|---|---|---|
Top 5 countries with best results | Average cases per day during prediction period | In or out of infection wave | Part of infection wave | Number of infection waves | |
1 | Pakistan | 2040 | In | Declining | 2 |
2 | Germany | 13,914 | In | Declining | 2 |
3 | Estonia | 507 | In | Declining | 2 |
4 | Slovakia | 2120 | In | Declining | 2 |
5 | United Kingdom | 37,476 | In | Declining | 3 |
Top 5 countries with worst results | Average cases per day during prediction period | In or out of infection wave | Part of infection wave | Number of infection waves | |
|---|---|---|---|---|---|
1 | Switzerland | 1884 | In | Declining | 2 |
2 | Croatia | 630 | In | Declining | 1 |
3 | Denmark | 869 | In | Declining | 2 |
4 | Georgia | 960 | In | Declining | 1 |
5 | Jordan | 943 | In | Declining | 1 |
