Page 148 - FULL FINAL HANDBOOK 20232024

Faculty of Science Handbook, Academic Session 2023/2024

parameters via the Expectation-Maximization (EM) block designs. Latin squares. Incomplete block designs.
algorithm. The Markov Chain Monte Carlo method. Factorial designs. Confounding. Fractional factorial
designs.
Assessment:
Continuous Assessment: 40% Assessment:
Final Examination: 60% Continuous Assessment: 40%
Final Examination: 60%
SIT3004 APPLIED STOCHASTIC
PROCESSES SIT3013 ANALYSIS OF FAILURE AND
SURVIVAL DATA
Time reversible Markov chains. Poisson processes.
Continuous-time Markov chains and birth and death Survival distributions, hazard models. Reliability of
processes. Brownian motion. Application to real-world systems, stochastic models. Censoring and life-tables.
phenomena, such as in finance. The product-limit estimator. Parametric survival models
under censoring. Cox proportional hazards model and
Assessment: other models with covariates.
Continuous Assessment: 40%
Final Examination: 60% Assessment:
Continuous Assessment: 40%
Final Examination: 60%
SIT3005 TIME SERIES AND
FORECASTING METHODS
Introduction to time series and forecasting. Time series SIT3015 INTRODUCTION TO
graphics. Simple forecasting methods. Transformation MULTIVARIATE ANALYSIS
and adjustments. Fitted values, residuals and prediction
intervals. Time series regression. Time series Matrix algebra and random vectors. Multivariate normal
decomposition. Exponential smoothing. ARIMA models. distribution. Wishart distribution and Hotelling
ARCH and GARCH models. distribution. Multivariate linear regression, canonical
correlation analysis. Dimensional reduction methods:
Assessment: principal component analysis, and linear discriminant
Continuous Assessment: 40% analysis. Clustering methods for unsupervised learning.
Final Examination: 60% Application of linear discriminant analysis, classification
and regression trees for supervised learning.
Assessment:
SIT3008 INTRODUCTION TO SURVEY Continuous Assessment: 40%
SAMPLING Final Examination: 60%
This course focuses on statistical sampling methods with
applications in the analysis of sample survey data. The
sampling methods include simple random sampling, SIT3016 GENERALIZED LINEAR MODELS
stratified random sampling, systematic sampling and
cluster sampling. Estimation of population parameters Introduction to generalized linear model based on the
for different sampling methods will be fully discussed. exponential family. For example, multiple linear
Special estimation techniques including ratio and regression for normal data, logistic regression for binary
regression estimations will be introduced in the context data, Poisson regression for counts, log linear for
of simple random sampling and stratified random contingency table, and gamma regression for continuous
sampling. Areas of application may include social non-normal data.
science and official statistics.
Study the theory of GLM including estimation and
Assessment: inference.
Continuous Assessment: 40%
Final Examination: 60% Introduction to fitting GLM in R.
Focus on the analysis of data: binary, count and
continuous, model selection, model evaluation,
SIT3009 STATISTICAL PROCESS interpretation, prediction and residual analysis.
CONTROL
Assessment:
Methods and philosophy of statistical process control. Continuous Assessment: 40%
Control charts for variables and attributes. Time- Final Examination: 60%
weighted control charts. Process capability analysis.
Multivariate control charts. Acceptance sampling plans.
Assessment: SIT3017 STATISTICAL LEARNING AND
Continuous Assessment: 40% DATA MINING
Final Examination: 60%
This course prepares students for applied work in data
science by building on students’ foundations of data
science skills. Students will learn advanced methods in
SIT3012 DESIGN AND ANALYSIS OF statistical learning and data mining, using appropriate
EXPERIMENTS computing tools such as R. The strengths of the diversity
of approaches are illustrated through analyses of real
Philosophy related to statistical designed experiments. world data sets covering commonly encountered data
Completely randomized one-factor design. Randomized types.
148