NeuralProphet: Explainable Forecasting at Scale

Abstract

We introduce NeuralProphet, a successor to Facebook Prophet, which set an industry standard for explainable, scalable, and user-friendly forecasting frameworks. With the proliferation of time series data, explainable forecasting remains a challenging task for business and operational decision making. Hybrid solutions are needed to bridge the gap between interpretable classical methods and scalable deep learning models. We view Prophet as a precursor to such a solution. However, Prophet lacks local context, which is essential for forecasting the near-term future and is challenging to extend due to its Stan backend.NeuralProphet is a hybrid forecasting framework based on PyTorch and trained with standard deep learning methods, making it easy for developers to extend the framework. Local context is introduced with auto-regression and covariate modules, which can be configured as classical linear regression or as Neural Networks. Otherwise, NeuralProphet retains the design philosophy of Prophet and provides the same basic model components.Our results demonstrate that NeuralProphet produces interpretable forecast components of equivalent or superior quality to Prophet on a set of generated time series. NeuralProphet outperforms Prophet on a diverse collection of real-world datasets. For short to medium-term forecasts, NeuralProphet improves forecast accuracy by 55 to 92 percent.

Long-Term Missing Value Imputation for Time Series Data Using Deep Neural Networks

“…It would also be interesting to study the proposed time-series imputation approach with other advanced feedforward neural networks such as convolutional neural network [45] and AR-NET [46]. Recently, a state-of-art time-series forecasting software NeuralProphet [47] combines the AR-NET and the Prophet software [48].…”

Experimental study of time series forecasting methods for groundwater level prediction

“…-NeuralProphet [26]. This is a neural network-based implementation of the Prophet model with some enhancements.…”

Real-Time Massive MIMO Channel Prediction: A Combination of Deep Learning and NeuralProphet

“…NeuralProphet: Within a short span of time, NP has emerged as a promising technique for different time-series prediction tasks [45]. In the context of DL, several time-series models, e.g., RNN and BiLSTM, as explained above, have been developed.…”

“…For the notational convenience, we will explain the model with D = 1, which will be later extended for multi-step, i.e., D future steps. In the context of channel prediction, the predicted value of NP for time instant t can be written as [45]…”

“…However, with regard to our application of channel prediction, we dropped a few as some of them are irrelevant for CSI prediction, e.g., holidays. For more details, the interested reader can refer to [45]. can be expressed as…”