Skip to content
VibeFormer

MODULE 22

Time Series Analysis

Stationarity, ACF/PACF, the ARIMA family, exponential smoothing, state-space models and deep forecasting.

15 lessons~7h reading

  1. 01

    Components of a Time Series

    BeginnerComing soon

    Trend, seasonality, cyclicity and noise; additive versus multiplicative models.

    24 min
  2. 02

    Stationarity

    IntermediateComing soon

    Strict and weak stationarity, why it matters for modelling, and differencing and transformation to achieve it.

    Assumes: Components of a Time Series

    28 min
  3. 03

    Unit Root Tests

    AdvancedComing soon

    ADF, KPSS and Phillips–Perron tests, their opposite null hypotheses, and interpreting conflicting results.

    Assumes: Stationarity · The Hypothesis Testing Framework

    26 min
  4. 04

    ACF and PACF

    IntermediateComing soon

    Autocorrelation and partial autocorrelation computed by hand, and reading order off the plots.

    Assumes: Stationarity

    30 min
  5. 05

    Autoregressive Models

    IntermediateComing soon

    AR(p) structure, the Yule–Walker equations, stationarity conditions and parameter estimation.

    Assumes: ACF and PACF

    28 min
  6. 06

    Moving Average Models

    IntermediateComing soon

    MA(q) structure, invertibility, and the duality between AR and MA representations.

    Assumes: Autoregressive Models

    26 min
  7. 07

    ARMA and ARIMA

    AdvancedComing soon

    Combining AR and MA, integrating for non-stationarity, and the Box–Jenkins model selection procedure.

    Assumes: Moving Average Models

    32 min
  8. 08

    Seasonal ARIMA and Exogenous Regressors

    AdvancedComing soon

    SARIMA notation, seasonal differencing, and SARIMAX with external drivers.

    Assumes: ARMA and ARIMA

    28 min
  9. 09

    Exponential Smoothing

    BeginnerComing soon

    Simple exponential smoothing, the smoothing constant, and worked forecast recursions.

    Assumes: Components of a Time Series

    26 min
  10. 10

    Holt and Holt–Winters

    IntermediateComing soon

    Adding trend and seasonal components, the ETS taxonomy, and damped trends.

    Assumes: Exponential Smoothing

    28 min
  11. 11

    Decomposition Methods

    IntermediateComing soon

    Classical decomposition, X-11, STL and their robustness to outliers.

    Assumes: Holt and Holt–Winters

    24 min
  12. 12

    Forecast Evaluation and Backtesting

    IntermediateComing soon

    MAE, RMSE, MAPE, sMAPE and MASE; rolling-origin cross-validation and why random k-fold is invalid here.

    Assumes: ARMA and ARIMA

    30 min
  13. 13

    Granger Causality and VAR

    AdvancedComing soon

    Vector autoregression, testing predictive causality, and the limits of the Granger notion.

    Assumes: Forecast Evaluation and Backtesting

    28 min
  14. 14

    State-Space Models and the Kalman Filter

    AdvancedComing soon

    The state-space formulation, the Kalman filter update derived, and a full numeric filtering pass.

    Assumes: Granger Causality and VAR · Hidden Markov Models

    34 min
  15. 15

    Deep Learning for Forecasting

    AdvancedComing soon

    Windowing for supervised learning, LSTM and TCN forecasters, N-BEATS, and transformer-based models.

    Assumes: Forecast Evaluation and Backtesting · Long Short-Term Memory

    30 min