'Why does the Prophet model also change the original dataset
I've been experimenting with the Prophet model from Facebook in Python. I find it to be quite efficient with forecasting my data. I just have one question though: I've tried to crop 120 rows from my dataset and predict the values for those 120 missing data.It works just fine but when I'm comparing the results I've found out that the values for the existing dates have changed. Why so ? Why does it not keep the values it has previously seen and try to predict only the missing values ?
Sources
This article follows the attribution requirements of Stack Overflow and is licensed under CC BY-SA 3.0.
Source: Stack Overflow
| Solution | Source |
|---|