Forecasting
See:
# Resources
- https://en.wikipedia.org/wiki/Forecasting
- Microsoft - Time Series Forecasting Best Practices & Examples
- Forecasting with a Time Series Model using Python:
- https://towardsdatascience.com/the-complete-guide-to-time-series-analysis-and-forecasting-70d476bfe775
- https://towardsdatascience.com/automl-for-time-series-advanced-approaches-with-fedot-framework-4f9d8ea3382c
- Predicting energy demand with neural networks
- Time series forecasting:
- Python Time Series Forecasting: A Practical Approach (WANDB)
- The Reasonable Effectiveness of Deep Learning for Time Series Forecasting
# Code
- #CODE
Skforecast
- https://joaquinamatrodrigo.github.io/skforecast/
- Package towards building Explainable Forecasting and Nowcasting Models with State-of-the-art Deep Neural Networks and Dynamic Factor Model on Time Series data sets with single line of code
- https://www.kdnuggets.com/2021/12/sota-explainable-forecasting-and-nowcasting.html
- Skforecast: time series forecasting with Python and Scikit-learn
- #CODE Pytorch-forecasting - Time series forecasting with PyTorch
- #CODE Deep_XF
- #CODE
statsforecast
- StatsForecast offers a collection of widely used univariate time series forecasting models, including autoARIMA and ETS modeling optimized for high performance using numba
- https://nixtla.github.io/statsforecast/
- #CODE
Darts
- https://unit8co.github.io/darts/
- Python library for easy manipulation and forecasting of time series. It contains a variety of models, from classics such as ARIMA, Prophet, deep neural networks (NBEATS, RNNs, Transformers)
- https://towardsdatascience.com/darts-swiss-knife-for-time-series-forecasting-in-python-f37bb74c126
- #CODE
Neuralforecast
- NeuralForecast is a Python library for time series forecasting with deep learning models. It includes benchmark datasets, data-loading utilities, evaluation functions, statistical tests, univariate model benchmarks and SOTA models implemented in PyTorch and PyTorchLightning
- https://nixtla.github.io/neuralforecast
- #CODE
Prophet (Facebook)
- https://facebook.github.io/prophet/
- Prophet is a procedure for forecasting time series data based on an additive model where non-linear trends are fit with yearly, weekly, and daily seasonality, plus holiday effects. It works best with time series that have strong seasonal effects and several seasons of historical data.
- https://research.fb.com/prophet-forecasting-at-scale/
- http://blog.fastforwardlabs.com/2017/03/22/prophet.html
- #CODE
NeuralProphet
- https://ourownstory.github.io/neural_prophet/
- A simple forecasting model based on Neural Networks in PyTorch
- #CODE
Hcrystalball
- Library that unifies the API for most commonly used libraries and modeling techniques for time-series forecasting in Python
- #CODE AtsPy: Automated Time Series Forecasting in Python
- #CODE
Greykite (Linkedin)
- https://linkedin.github.io/greykite/
- A flexible, intuitive and fast forecasting library
- #CODE Scalecast
# Talks
# Books
# References
- #PAPER Time Series Forecasting With Deep Learning: A Survey (Lim 2020)
- #PAPER N-BEATS: Neural basis expansion analysis for interpretable time series forecasting (Oreshkin 2020)
- #PAPER A flexible forecasting model for production systems (Hosseini 2021)
- #PAPER An Experimental Review on Deep Learning Architectures for Time Series Forecasting (Lara-Benitez 2021)
- #PAPER Temporal Fusion Transformers for interpretable multi-horizon time series forecasting (Lim 2021)
- #PAPER
Understanding forecast reconciliation (Hollyman 2021)
- Forecast Reconciliation - a process by which independently generated forecasts of a collection of linearly related time series are reconciled via the introduction of accounting aggregations that naturally apply to the data
- #PAPER A Review on Deep Sequential Models for Forecasting Time Series Data (Mahdi Ahmed 2022)
- #PAPER
DeepTime: Deep Time-Index Meta-Learning for Non-Stationary Time-Series Forecasting (Woo 2022)
- #CODE https://github.com/salesforce/DeepTime
- DeepTime is a deep time-index based model trained via a meta-learning formulation, yielding a strong method for non-stationary time-series forecasting
- https://pub.towardsai.net/meta-learning-for-time-series-forecasting-deeptime-in-pytorch-lightning-3d97351099c6
- #PAPER Forecast evaluation for data scientists: common pitfalls and best practices (Hewamalage 2022)
- #PAPER #REVIEW Review of ML and AutoML Solutions to Forecast Time-Series Data (Alsharef 2022)