Prophet VS darts

My rescue came from discovering some of the alternatives to ARIMA/SARIMA, which was the extent of what we had covered for time series data. A series of searches eventually led me to some automated time series analysis packages, one of which was Prophet, an open source time series package released by Facebook's core data science team. This was a life saver, being a much more efficient and more effective forecasting tool than sloooowly iterating through ARIMA/SARIMA models that seemed to want to fight with me. If you're going to do a time series analysis for your capstone, I strongly suggest taking a look at using Prophet.

Prophet

Prophet

Prophet: Automação preditiva.

Prophet - It is a time-series forecasting library built by Facebook. 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. Prophet is robust to missing data and shifts in the trend, and typically handles outliers well. Link - https://github.com/facebook/prophet

We run an anomaly detection app powered by Facebook's Prophet forecasting library. It tells us if metrics dip or rise in unexpected ways ("Did signups drop? Is something broken with that flow?"). We built the service because customers kept reaching out to tell us some feature broke before we noticed. Normally these issues show up in product data, so the app looks for these anomalies and tells us when they happen.

active support

Darts gives you a lot of options, including newer deep learning approaches like NBEATS and NHiTS.

https://unit8co.github.io/darts/

I would recommend Darts in Python [1]. It's easy to use (think fit()/predict()) and includes

* Statistical models (ETS, (V)ARIMA(X), etc)

* ML models (sklearn models, LGBM, etc)

* Many recent deep learning models (N-BEATS, TFT, etc)

* Seamlessly works on multi-dimensional series

* Models can be trained on multiple series

* Many models offer rich support for probabilistic forecasts

* Model evaluation is easy: Darts has many metrics, offers backtest etc

* Deep learning scales to large datasets, using GPUs, TPUs, etc

* There's even now an explainability module for some of the models - showing you what matters for computing the forecasts

* (coming soon): an anomaly detection module :)

* (also, it even include FB Prophet if you really want to use it)

Warning: I'm probably biased because I'm Darts creator.

[1] https://github.com/unit8co/darts

To be fair, Darts looks pretty good relative to forecast: https://github.com/unit8co/darts

I would generally prefer R for this kind of stuff as the experts generally write the code, but Darts seems OK and is well-tested, at the very least (haven't had a chance to use it in anger yet).

There is also a version of N-BEATS in Darts (https://github.com/unit8co/darts) that extends the original N-BEATS by * Accepting exogenous covariate time series * Being able to produce probabilistic forecasts * Working on multivariate time series (all of this out of the box, fit() / predict() style) :D

Have you checked out darts in python for time series analysis? https://unit8co.github.io/darts/

N-HiTS, N-BEATS, as well as any sklearn model (or LightGBM, which often performs very well) are all available in Darts, with the same easy & simple API: https://github.com/unit8co/darts

DARTS https://github.com/unit8co/darts (PyTorch based)