ARIMA and Exponential Smoothing Explained With Real Numbers

Two classic forecasting families, ARIMA and exponential smoothing, built from the ground up with hand-worked numbers. We cover differencing and the ADF test, AR, MA and ARMA models, Yule-Walker by hand, the lag operator, reading ACF and PACF plots, residual checks with Ljung-Box, simple, Holt and Holt-Winters smoothing, seasonal ARIMA, and a rolling-origin evaluation in Python with statsmodels. This is lesson 46 of Machine Learning from Zero to Advanced. It builds on lesson 25 on time series basics. Subscribe so you do not miss lesson 47 on weight initialization and optimizers. 0:00 Intro 0:41 Stationarity and Differencing 2:37 Autoregressive Models 4:56 Moving Average and ARMA 6:34 Identifying and Fitting 8:44 Exponential Smoothing 12:01 Seasonal ARIMA 12:31 Python and Evaluation 13:58 Comparing With ML Models 14:53 Recap #ARIMA #ExponentialSmoothing #TimeSeries #MachineLearning #Forecasting