Apr-23-2020, 05:32 PM
Your dataset has X and y variables in it, say days and CV-19 cases. These are all known values.
You send it to TTS, then train on the train values. Once trained, you can validate your model by testing the model against the test set. So, when it predicts that day 27 you should have 500 cases and the actual value is 600, you have an error of 100. You can then get statistics (typically the mse - mean squared error) on how close your model actually predicts reality.
Once you are comfortable with your model you can use it to predict new values - what will we be looking at on day 300?
You send it to TTS, then train on the train values. Once trained, you can validate your model by testing the model against the test set. So, when it predicts that day 27 you should have 500 cases and the actual value is 600, you have an error of 100. You can then get statistics (typically the mse - mean squared error) on how close your model actually predicts reality.
Once you are comfortable with your model you can use it to predict new values - what will we be looking at on day 300?