Feature engineering efforts mainly have two goals: Creating the correct input dataset to feed the ML algorithm: In this case, the purpose of feature engineering in time series forecasting is to create input features from historical row data and shape the dataset as a supervised learning problem .5 Eki 2021