Train-Test-Split + define outputs needed from prediction models

[harkar1978]

Train-Test Split completed. Metrics required to evaluate different prediction models in progress

[AnnaWey]

MAPE as metric for our models to make it more comparable across models