.. image:: pyeco.png :height: 20 :alt: Economie :target: http://www.xavierdupre.fr/app/ensae_teaching_cs/helpsphinx/td_2a_notions.html#pour-un-profil-plutot-economiste .. image:: pystat.png :height: 20 :alt: Statistique :target: http://www.xavierdupre.fr/app/ensae_teaching_cs/helpsphinx/td_2a_notions.html#pour-un-profil-plutot-data-scientist Timeseries - Séries temporelles +++++++++++++++++++++++++++++++ *Notebooks* .. toctree:: :maxdepth: 2 ../notebooks/ml_timeseries_base ../notebooks/td2a_timeseries ../notebooks/td2a_timeseries_correction ../notebooks/seasonal_timeseries (à venir : modèles SETAR pour les séries non périodiques, modèles proies prédateurs) Quelques rappels sur les lissages, saisonnalités, modèles linéaires appliqués aux séries temporelles : `Projet Machine Learning pour la Prévision: séries temporelles `_ *Lectures* * `Time series analysis with pandas `_ * `Consistent Algorithms for Clustering Time Series `_ * `Learning Time Series Detection Models from Temporally Imprecise Labels `_ * `Time Series Prediction With Deep Learning in Keras `_ * `Sequence Classification with LSTM Recurrent Neural Networks in Python with Keras `_ (voir `LSTM `_) * `Time Series Classification and Clustering with Python `_ * `Dynamic Time Warping `_ * `Functional responses, functional covariates and the concurrent model `_ * `Fast and Accurate Time Series Classification with WEASEL `_ (text and timeseries) * `Forecasting at Scale `_ (Facebook) * `SETAR `_ : prédiction sur des modèles en apparence cycliques mais non périodiques (type proies-prédateurs, chaotiques), SETAR = Self-Exciting Threshold AutoRegressive * `Using predator-prey models on the Canadian lynx series `_, `Inference for nonlinear dynamical systems `_ * `Milestones of Deep Learning `_ * `Engineering Extreme Event Forecasting at Uber with Recurrent Neural Networks `_ * `Holt-Winters seasonal method `_, `Initializing the Holt-Winters method `_ * `A Comparison of Estimation Methods for Vector Autoregressive Moving-Average Models `_ * `A state space framework for automatic forecasting using exponential smoothing methods `_ * `Clustering Time Series with Nonlinear Dynamics: A Bayesian Non-Parametric and Particle-Based Approach `_ *Compétitions* * Kaggle Web Traffic Time Series Forecasting `code `_, `modèle `_ *Modules* *`merlion `_ * `statsmodels `_ * `sktime `_, suit l'API de scikit-learn, assez complet * `pyflux `_ (la documentation est plutôt bien faite) * `fbprophet `_ (requires `pystan `_) *autres* * `pmdarima `_, ARIMA, saisonnalités, pipelines * `kats `_, modèle SARIMA, clustering de séries temporelles * `pastas `_, pour les séries hydrologiques * `darts `_ * `orbits `_, bayesian forecasting * `tensorflow `_ * `Rob J Hyndman software `_ (disponible uniquement en R) * `influxdb `_ (An Open-Source Time Series Database) * `seasonal `_ * `seglearn `_ * `flow-forecast `_ * `Auto_TS `_, automated learning for time series Un peu plus expérimental : * `sulekha_holtwinters `_, Holt-Winters sur :epkg:`Spark` * `holtwinters.py `_, Holt-Winters dans un fichier.