Implement a jupyter notebook with a full guide of 'how-to-timeseries in FEDOT'
Created by: J3FALL
Fedot provides functions for timeseries forecasting and some examples can be seen in examples/time_series_forecasting.py.
It could be useful to extend the examples with a 'comprehensive guide for dummies' how to forecast timeseries using Fedot framework.
What can be included in:
- Timeseries format representations? Simple visualization. Fedot TS-format
- Synthetic timeseries generation
- Some real examples? (Stock prices for instance)
- Onestep prediction with 1 model (without exog)
- What is exog values? Onestep with exog
- Multistep predictions with/without exog
- Multiscale predictions (trend + residual)
- Predictions with composite model
- Composer-based chain generation for timeseries (?)