Statistical Marketing Tools
Unit code: HMS794
|Credit points||12.5 Credit Points|
|Duration||One semester / teaching period|
|Contact hours||3 Hours per Week|
|Prerequisites||HMS780 Multivariate Statistics Multivariate Statistics|
Related course(s)A unit of study in the Master of Science (Applied Statistics).
Aims and objectives
- Demonstrate an understanding of the basic theory and principles of data miningPerform and interpret association analyses including market segmentation and market basket analysis.
- Apply prediction and classification methods such as regression, neural networks and decision trees, understanding how to choose between these methods.
- Use quantitative analysis techniques commonly used in market research; mapping techniques, including multidimensional scaling and correspondence analysis, and preference techniques, including conjoint analysis.
- Read, understand and critically assess research publications using data mining methods and statistical tools for marketing
Teaching methodsThe classes are held in a computer laboratory and practical exercises are integrated with class teaching throughout the sessions. SAS software will be supplied by DVD. Please note: SAS will only run on a Windows platform.
Market research analysts in both commerce and industry make daily decisions regarding their products and services in today's complex and competitive markets. These decisions, important to the welfare of their companies, must be based on the best available information, usually gathered through surveys. The methods included will be selected from the following with possible changes in order:
- Mapping techniques, including multidimensional scaling and correspondence analysis
- Preference techniques, including conjoint analysis
- Market segmentation methods for finding statistically significant groups in data using methods, such as cluster analysis and self organizing maps
- Data Mining approaches and procedures using appropriate software.
- Association using market basket analysis and link analysis.
- Classification using logistic regression, classification trees, neural networks and memory based reasoning.
- Prediction using regression, regression trees, neural networks and memory based reasoning.
Week 1 Multidimensional scaling 1.
Week 2 Multidimensional scaling 2.
Week 3 Correspondence analysis 1.
Week 4 Correspondence analysis 2.
Week 5 Conjoint Analysis 1.
Week 6 Conjoint Analysis 2.
Week 7 Intro to data mining and EDA
Week 8 Market basket analysis and classification trees.
Week 9 Logistic regression, regression and regression trees.
Week 10 Neural networks for classification and prediction.
Week 11 Cluster analysis and self organising maps.
Week 12 Student Presentations