MATLAB and Simulink Training

Time-Series Modeling in MATLAB

Course Details

This one-day course provides a comprehensive introduction to time-series modeling using MATLAB® and Econometrics Toolbox™. The course is intended for economists, analysts and other financial professionals with prior experience of MATLAB who require to develop and maintain time-series models. The course is designed to follow the standard Box-Jenkins procedure for developing time-series models.

High-level course themes include:

  • Preprocessing time-series data
  • Identifying long-term and seasonal trends in time-series data
  • Testing data stationarity using hypothesis tests
  • Creating and fitting ARIMA and GARCH time-series models to a data set
  • Comparing different model fits for the same data
  • Analyzing model dynamics using Monte Carlo simulations
  • Forecasting data using fitted models

Day 1 of 1


Preparing Data for Model Fitting

Objective: Prepare time-series data for model fitting by identifying trends and applying data transformations.

  • Removing exponential trends
  • Identifying polynomial and seasonal trends
  • Testing for data stationarity
  • Stationarizing data
  • Unit-root tests

Model Selection and Fitting

Objective: Use diagnostic tools to select a group of suitable candidate ARIMA and GARCH models for a given time series. Identify, create and fit candidate time-series models to data.

  • Computing autocorrelation and partial autocorrelation
  • Selecting models using formal tests
  • Selecting candidate ARIMA and GARCH models for a given data set
  • Creating and fitting time series models to a data set

Evaluating Model Appropriateness

Objective: Compute and evaluate model diagnostics to ensure model correctness, suitability and reliability.

  • Inferring model residuals
  • Testing residuals for Normality
  • Analyzing model diagnostics and goodness-of-fit statistics
  • Evaluating significance of individual model terms
  • Comparing models

Forecasting and Simulation

Objective: Forecast models to predict future data. Simulate sample trajectories and statistics by applying Monte Carlo simulation techniques.

  • Forecasting data using fitted models
  • Using in-sample forecasts to evaluate model appropriateness
  • Monte Carlo model simulation
  • Backtesting models

Level: Intermediate

Prerequisites:

This program has been approved by GARP and qualifies for 7 GARP CPD credit hours. If you are a Certified FRM or ERP, please record this activity in your credit tracker at https://www.garp.org/cpd

Duration: 1 day