High-Quality Time-Series Analysis
Tshepo Chris Nokeri · Apress eBooks · 2021
The preceding chapter covered seasonal ARIMA. After all the considerable effort in data preprocessing and hyperparameter optimization, the model generates considerable errors when forecasting future instances of the series. For a fast and automated forecasting procedure, use Facebook’s Prophet; it forecasts time-series data based on nonlinear trends with seasonality and holiday effects. This chapter introduces Prophet and presents a way of developing and testing an additive model. First, it discusses the crucial difference between the Statsmodels package and the Prophet package.