Predictive Analytics investigates multiple business challenges and cases, including segmentation, brand positioning, product choice modelling, pricing research, finance, sports, Web and text analytics, and social network analysis. Using techniques such as cross-sectional data, time series, spatial, and even spatiotemporal data, this module links the technological, scientific and business components together to help the student understand the complex nature of modelling in a business context. Using Python and ‘R’, students will build models to explore different business questions.
Properties of Statistical Distributions: Moment Generating Functions, Cumulant Generating Function, Chebychev and Markov Inequalities, Edgeworth and Cornish-Fischer expansions, Copula functions.
Matric Relationships: Matrix Sweep and Partial Correlation, Singular Value Decomposition, Spectral, Conjugate and Cholesky Decomposition
Linear and Non-Linear Modeling: Maximum Likelihood Estimators, Fischer’s Analysis, PCR, Factor Analysis and PLSR. Segmentation and Tree Models, Additive Models,
Model Goodness Measures: Goodness-of-fit tests, Kolmogorov-Smirnov (KS) Statistic, AIC and SBC, Hosmer-Lemeshow goodness-of-fit Test
The module is delivered through a combination of lectures, and the tutor leads IT labs. Additionally, various e-learning aids are used.
-
Lectures: Teach fundamental concepts of predictive analytics and real-world applications.
-
Hands-on labs: Provide practical experience with real-world datasets.
-
Case studies: Use real-world examples to illustrate the impact of predictive analytics.
-
Collaborative projects: Encourage group projects to develop and implement predictive analytics solutions.
-
Self-directed learning: Provide access to online resources and readings for self-directed learning.
-
Assessments: Assess students' progress and provide feedback.
| Module Content & Assessment | |
|---|---|
| Assessment Breakdown | % |
| Other Assessment(s) | 100 |