[docs] Outdated tutorials

[leoniewgnr]

Current structure

Tutorials

Trend: can be removed, as fully covered in tutorial 2

Seasonality: can be removed, only multiplicative seasonality should be covered seasonality_mode='multiplicative'

Auto-Regression: super outdated and should definitely be removed, only thing that could be covered in another tutorial is m.highlight_nth_step_ahead_of_each_forecast(step_number=m.n_forecasts)

Lagged regressors: can be removed, as fully covered in tutorial 5

Events: can be removed as covered in Tutorial 7

Future regressors: can be removed as covered in Tutorial 6

Feature Guides

would propose to delete them completely and build new ones upon them (see new structure)

Proposed new structure

• Delete old tutorials completely

Tutorials

• Basic tutorials: Tutorials 01 - 11
• Feature guides:
  • autoregression_yosemite_temps.ipynb --> delete as covered in basic tutorials
  • collect_predictions.ipynb --> keep as it is
  • conditional_seasonality_peyton.ipynb --> publish
  • events_holidays_peyton_manning.ipynb --> delete as covered in basic tutorials
  • global_local_modeling.ipynb --> publish, but needs a revamp
  • global_modeling.ipynb --> delete as covered in global local modeling
  • lagged_covariates_energy_ercot.ipynb --> delete as covered in basic tutorials
  • Live_plot_during_training.ipynb --> do not publish as not working
  • migration_from_prophet.ipynb --> publish (if possible, could be bugged)
  • network_architecture_visualization.ipynb --> publish
  • plotly.ipynb --> publish
  • prophet_to_torch_prophet.ipynb --> delete as covered in migration_from_prophet
  • season_multiplicative_air_travel.ipynb --> publish, maybe extend to other multiplicative things
  • sparse_autoregression_yosemite_temps.ipynb --> publish as it is a continuation of tutorial 04: Autoregression
  • sub_daily_data_yosemite_temps.ipynb --> publish
  • test_and_crossvalidate.ipynb --> publish, but could be improved in the future
  • trend_peyton_manning.ipynb --> delete as covered in basic tutorials
  • uncertainty_conformal_prediction.ipynb --> waiting for answer of https://github.com/ourownstory/neural_prophet/pull/1225
  • uncertainty_quantile_regression.ipynb --> waiting for answer of https://github.com/ourownstory/neural_prophet/pull/1225
  • Regularization would be nice if added
  • Multi-step-forecasting (esp. plotting) would be nice if added
• Hyperparameter selection guide
• Application examples: stay the same

The Science behind

• Model overview
• presentation
• Changes from Prophet

[noxan]

@leoniewgnr is this one done with your latest PR?

[leoniewgnr]

yes is done