Third Year Modules

 

STAT321: Linear Models and Time Series Analysis

  Stream: BCom: Statistics
  Credits: 30
  Pre-requisites: STAS211 (or STAS201) and STAS202
 
 Module Content
  • Specification, estimation and testing of single-equation and multi-equation linear models.
  • Monte Carlo simulation.
  • Distributed and Koyck lag models.
  • Tests for simultaneity and homogeneity.
  • Estimation using OLS, GLS, WLS, FGLS and SUR.
  • Properties of stochastic time series models.
  • Extrapolation and smoothing.
  • Moving averages, exponential smoothing and Holt-Winters model, regression models and autoregressive error models.
  • The Box-Jenkins approach: ARMA and ARIMA models.

 

 

 

STAT311: Advanced Statistical Inference

  Stream: BSc - Mathematical Statistics
  Credits: 30
  Pre-requisites:  STAS201 and STAS202
 
Module Content
  • Theory of limiting distributions.
  • Theory and application of sampling distributions.
  • Theory and application method of moments and maximum likelihood estimation.
  • Estimator evaluation using statistical properties.
  • Large sample properties of estimators.
  • Most powerful and UMP tests.
  • Generalised likelihood ratio tests.
  • Neyman-Pearson lemma.
  • Derivation and application of hypothesis tests.
  • Introduction to Bayesian inference.
  • Specification, estimation and tests of single-equation and multi-equation linear models.
  • Tests for simultaneity and homogeneity in linear models.
  • Estimation of linear models using OLS, GLS, WLS and FGLS.
  • Monte Carlo simulation.

 

 

 

STAT312: Advanced Data Analytics

  Stream: BCom – Statistics & BSc – Mathematical Statistics
  Credits: 30
  Pre-requisites: STAS201 and STAS202
 
Module Content
  • Data collection and wrangling.
  • Cross-validation methods (holdout and k-fold).
  • Linear regression in R.
  • Analysis of variance (ANOVA).
  • Logistic regression.
  • Non-parametric regression (knn and Nadaraya-Watson).
  • Classification (knn, discriminant analysis and k-means clustering).
  • Model assessment techniques (accuracy, sensitivity, and specificity).
  • Non-parametric hypothesis tests.
  • Principal component analysis.