| Code | Faculty | Department |
|---|---|---|
| 02240193 | Faculty of Natural and Agricultural Sciences | Department: Statistics |
| Credits | Duration | NQF level |
|---|---|---|
| Minimum duration of study: 1 year | Total credits: 135 | NQF level: 08 |
Note: Additional modules may be required in order to reach the desired level of competency
Minimum credits: 135
Core credits: 75
Elective credits: 60 (choose four electives from the list)
A student cannot get credit for either RAL 780 or MET 720 with a WST undergraduate major.
Module content:
The emphasis is on the theoretical understanding and practical application of advances in statistical modelling. The following topics are covered: Single equation models: Nonparametric regression. Bootstrap procedures within regression analysis, k-nearest neighbour classification. Modelling categorical dependent variables - Logit/Probit models. Multiple outputs. Linear regression of an indicator matrix. Ridge regression. Non-linear regression modelling. Some new developments in regression and classification.
Simultaneous equation models: Specification, identification and estimation of simultaneous equation models.
Module content:
Point and Interval estimation. Sampling distributions, central limit theorem, simulations and Bootstrap. Bayesian inference, posterior distribution. Hypotheses testing using confidence intervals, ratio tests, simulated null distributions and power function. A student cannot get credit for this module with a WST undergraduate major.
Module content:
Matrix methods in statistics. Simple and multiple regression models. Sums of squares of linear sets. Generalised t- and F-tests. Residual analysis. Diagnostics for leverage, influence and multicolinearity. Indicator variables. Regression approach to analysis of variance. Weighted least squares. Theory is combined with practical work. A student cannot get credit for this module with a WST undergraduate major.
Module content:
Refer to the document: Criteria for the research management process and the assessment of the honours essays, available on the web: www.up.ac.za under the Department of Statistics: Postgraduate study.
Module content:
A compulsory bootcamp must be attended as part of this module – usually presented during the last week of January each year (details are made available by the department ). The bootcamp will cover the basics of research to prepare students for the research component of their degree. The bootcamp should be done in the same year as registration for STK 795/WST 795. Each year of registration for the honours degree will also require the attendance of three departmental seminars. Students should ensure that their attendance is recorded by the postgraduate co-ordinator present at the seminars. The department approves the seminars attended. In addition, students are required to present their STK 795/WST 795 research in the department during the year of registration for these modules.
Module content:
Mixtures of distributions and regressions, frequentist and Bayes estimation. Latent components, soft allocation and belongings. Applications in unstructured data, including text data. Identification and interpretation of behavioural patterns.
Module content:
Simple random sampling. Estimation of proportions and sample sizes. Stratified random sampling. Ratio and regression estimators. Systematic and cluster sampling. Introduction to spatial statistics. Spatial sampling – both model and design based approaches.
Module content:
Quality control and improvement. Shewhart, cumulative sum (CUSUM), exponentially weighted moving average (EWMA) and Q control charts. Univariate and multivariate control charts. Determining process and measurement systems capability. Parametric and nonparametric (distribution-free) control charts. Constructing control charts using Microsoft Excel and/or SAS. Obtaining run-length characteristics via simulations, the integral equation approach, other approximate methods and the Markov-chain approach.
Module content:
Efficient programming, Monte Carlo simulation, sampling of discrete and continuous probability models, General transformation methods, Accept-reject methods, Monte Carlo integration, importance sampling, numerical optimisation, Metropolis-Hastings algorithm, GIBBS sampling.
Module content:
This module considers specific topics from the diverse field of statistics as deemed supportive towards the training of the cohort of scholars.
Module content:
Specification of linear mixed model, model assumptions, estimation (REML and ML), diagnostics, hypothesis tests, interpretation of parameter estimates, calculating predicted values. Specific models: two- and three-level models for clustered data, intraclass correlation coefficients, repeated measures data, random coefficient models for longitudinal data, models for clustered longitudinal data, models for data with crossed random factors. Using statistical software to analyse LMMs.
Copyright © University of Pretoria 2024. All rights reserved.
Virtual Campus
Get Social With Us
Download the UP Mobile App