Skip to content

Latest commit

 

History

History
52 lines (46 loc) · 2.64 KB

TODO.md

File metadata and controls

52 lines (46 loc) · 2.64 KB
Raw

Basic Tenants

  • Ease of Use > Accuracy > Speed (with speed more important with 'fast' selections)
  • Availability of models which share information among series
  • All models should be probabilistic (upper/lower forecasts)
  • New transformations should be applicable to many datasets and models
  • New models need only be sometimes applicable
  • Fault tolerance: it is perfectly acceptable for model parameters to fail on some datasets, the higher level API will pass over and use others.
  • Missing data tolerance: large chunks of data can be missing and model will still produce reasonable results (although lower quality than if data is available)

Assumptions on Data

  • Series will largely be consistent in period, or at least up-sampled to regular intervals
  • The most recent data will generally be the most important
  • Forecasts are desired for the future immediately following the most recent data.
  • trimmed_mean to AverageValueNaive

0.6.14 🇺🇦 🇺🇦 🇺🇦

  • prevent excessive use of 'CenterSplit' and other macro_micro style transformers
  • added ElasticNetwork as subsidiary regression model option
  • KalmanSmoothing, BKBandpassFilter added on_inverse option
  • add threshold arg to AlignLastValue

Unstable Upstream Pacakges (those that are frequently broken by maintainers)

  • Pytorch-Forecasting
  • Neural Prophet
  • GluonTS

New Model Checklist:

* Add to ModelMonster in auto_model.py
* add to appropriate model_lists: all, recombination_approved if so, no_shared if so
* add to model table in extended_tutorial.md (most columns here have an equivalent model_list)
* if model has regressors, make sure it meets Simulation Forecasting needs (method=="regressor", fails on no regressor if "User")

New Transformer Checklist:

* Make sure that if it modifies the size (more/fewer columns or rows) it returns pd.DataFrame with proper index/columns
* add to transformer_dict
* add to trans_dict or have_params or external
* add to shared_trans if so
* oddities_list for those with forecast/original transform difference
* add to docstring of GeneralTransformer
* add to dictionary by type: filter, scaler, transformer
* add to test_transform call

New Metric Checklist:

* Create function in metrics.py
* Add to mode base full_metric_evaluation  (benchmark to make sure it is still fast)
* Add to concat in TemplateWizard (if per_series metrics will be used)
* Add to concat in TemplateEvalObject (if per_series metrics will be used)
* Add to generate_score
* Add to generate_score_per_series (if per_series metrics will be used)
* Add to validation_aggregation
* Update test_metrics results
* metric_weighting in AutoTS, get_new_params, prod example, test