Complete temporal forecasting track: from EDA and stationarity tests to classical ARIMA/SARIMA, ML feature engineering with XGBoost, recurrent deep learning (RNN/LSTM/GRU), and production walk-forward validation.
Complete all 5 milestones to unlock and download your verified certificate (5 remaining).
Understand stationarity, trend-seasonality decomposition (STL), autocorrelation (ACF/PACF), and temporal data cleaning.
Mathematical foundations of AR, MA, ARMA, ARIMA, seasonal SARIMA, Box-Jenkins methodology, and residual diagnostics.
Transform temporal data into supervised ML formats: lag features, rolling statistics, leak-safe cross-validation, and XGBoost/LightGBM.
Vanishing gradients, gated recurrent units (GRUs), Long Short-Term Memory (LSTM), seq2seq architectures, and sequence forecasting.
Time-series Transformers, Prophet, multi-step forecasting, walk-forward validation, drift monitoring, and production MLOps.