The aims of the module are to:To instill an understanding of the foundations for advanced analytics and be able to apply the techniques underpinning the machine learning processes.
Data Pre-Processing and Exploration
Preparing the data (data transformation); missing data (removal and impute data)Outliers; correlations; sparse data; normalisation and standardisation; attribute selection; Dimensionality and data reduction
Machine Learning Techniques
Classification Techniques (detailed walk through of particular algorithms): e.g. Decision Trees: C4.5; Lazy Learners kNN; Bayes Classifiers, Linear Regression, Logistic Regression Regression models; VotingEnsemble Methods: Bagging, Boosting, AdaBoost, Random Forests; Clustering Techniques:Partitioning Methods e.g. kMeans Clustering; Agglomerative Vs Divisive Hierachical Clustering e.g. BIRCH Density Methods e.g. DBSCAN; Association Rules Discovery: e.g. APriori Algorithm
Model Evaluation and Predicting Performance Techniques
Test Options: Holdout Method, Stratification, Cross-validation - 10FoldCrossValidation, Sampling with Replacement(Bootstrap); Confusion Matrix; Error RateAccuracy; Under and over fitting, generalisabilitySensitivity(Recall) and Specificity; Precision; F Measure; ROC; Root mean square; Lift; Model comparison(T-test (student and Welch’s), ANOVA, Binomial distribution)
| Module Content & Assessment | |
|---|---|
| Assessment Breakdown | % |
| Other Assessment(s) | 50 |
| Formal Examination | 50 |