Advanced Analytics in a Big Data World
Business Knowledge Series course
Presented by Bart Baesens, Ph.D. Professor at KU Leuven (Belgium), and lecturer at the University of Southampton (UK); or Christophe Mues, Ph.D., Professor at the School of Management of the University of Southampton (UK); or Cristian Bravo, Ph.D., Associate Professor, Business Analytics, University of Southampton (UK); or Wouter Verbeke, Ph.D., Assistant Professor, Business Informatics, University of Brussels (Belgium); or Stefan Lessmann, Ph.D., Professor, School of Business and Economics, Humboldt University (Germany)
In today's big data world, many companies have gathered huge amounts of customer data about marketing success, use of financial services, online usage, and even fraud behavior. Given recent trends and needs such as mass customization, personalization, Web 2.0, one-to-one marketing, risk management, and fraud detection, it becomes increasingly important to extract, understand, and exploit analytical patterns of customer behavior and strategic intelligence. This course helps clarify how to successfully adopt recently proposed state-of-the art analytical and data science techniques for advanced customer intelligence applications. This highly interactive course provides a sound mix of both theoretical and technical insights as well as practical implementation details and is illustrated by several real-life cases. References to background material such as selected papers, tutorials, and guidelines are also provided.
Learn how to
- Apply a series of powerful, recently developed, cutting-edge analytical and data science techniques.
- Ensure the practical application of these techniques to optimize strategic business processes and decision making.
- Explore a futuristic vision of how emerging data science techniques might change your key business processes.
- Deploy, monitor, and optimally backtest analytical models.
Who should attend
Those involved in estimating, monitoring, auditing, or maintaining models for various types of customer intelligence; those involved with using data mining techniques for various types of customer intelligence, job titles including business analysts in various settings (for example, risk management, manufacturing, telco, retail, advertising, public, pharmaceutical, and so on), marketing/CRM managers, fraud managers, customer intelligence managers, risk analysts, CRM analysts, marketing analysts, senior data analysts, and data miners
Before attending this course, you should know how to:
- Preprocess data (such as sampling, missing values, outliers, categorization, and so on).
- Develop predictive models using logistic regression.
- Develop predictive models using decision trees.
- Develop descriptive models using basic segmentation techniques.
- Quantify the performance of predictive models (such as lift curves, ROC curves, and so on).
This course addresses SAS Enterprise Miner software.
Refresher: The Analytics Process Model
Refresher: Decision Trees
- Basic nomenclature (definition of customer, definition of target, and so on).
- Data collection and preprocessing (sampling, missing values, outliers, weights of evidence, and so on).
- Predictive versus descriptive analytics (data mining).
- Putting analytics to work.
- Analytic model requirements (performance, interpretability, operational efficiency, compliance).
- Key application areas (CRM, risk management, fraud, online analytics).
- Splitting/stopping/assignment decision.
- Key algorithms: C4.5 (See5), CART, CHAID.
- Recommendations for using decision trees in a business context.
- Splitting/stopping/assignment criteria.
- Case study: using regression trees for loss forecasting.
Alternative Rule Representation Formats
- Random forests.
- Rule types (propositional, oblique, M-of-N, fuzzy, and so on).
- Decision tables (lexicographical ordering, contraction methods, and so on).
- Decision diagrams.
- Case study: decision tables and diagrams for customer scoring.
- Case study: decision tables for textual knowledge verification.
Support Vector Machines (SVMs)
- Multilayer perceptrons (MLPs).
- MLP types (RBF, recurrent, and so on).
- Weight learning (backpropagation, conjugate gradient, and so on).
- Overfitting, early stopping, and weight regularization.
- Architecture selection (grid search, SNC, and so on).
- Input selection (Hinton graphs, likelihood statistics, brute force, and so on).
- Self-organizing maps (SOMs) for clustering.
- Case study: using SOMs for country corruption analysis.
Opening Up the Neural Network and SVM Black Box
- Linear programming.
- The kernel trick and Mercer theorem.
- SVMs for classification and regression.
- Multiclass SVMs (one-versus-one, one-versus-all coding).
- Hyperparameter tuning using cross-validation methods.
- Case study: benchmarking SVM classifiers.
Bayesian Network Classifiers
- Business applications of neural networks and SVMs.
- Rule extraction methods (pedagogical versus decompositional approaches such as neurorule, neurolinear, and trepan).
- Two-stage models (combining white-box linear models with black-box high-performing neural networks).
- Case studies.
- Naive Bayes.
- Tree augmented Naive Bayes (TAN).
- Unrestricted Bayesian network classifiers.
- Bayesian inference.
- Case study: Bayesian networks for churn prediction.
Social Network Learning and Inference
- Survival probabilities versus hazard rates.
- Kaplan Meier analysis.
- Parametric survival analysis.
- Proportional hazards regression.
- Time varying covariates.
- Competing risks.
- Neural networks for survival analysis.
- Case study: survival analysis for Customer Lifetime Value (CLV) modeling.
Monitoring and Backtesting Analytical Models
- Implicit versus explicit social networks.
- Learning using networked data.
- Key application areas (Facebook/Twitter, churn, fraud, online analytics, and so on).
- Markov random fields.
- Homophily (guilt by association).
- Local classifiers.
- Relational classifiers (relational neighbor, probabilistic relational neighbor, relational logistic regression, and so on).
- Collective inference (Gibbs sampling, iterative classification, and so on).
- Case study: using social networks for churn detection in a telco context.
Other Learning Algorithms and Applications (Short)
- Quantitative versus qualitative model monitoring.
- Model backtesting (model stability, model discrimination, model calibration, binomial/Hosmer-Lemeshow test, traffic light indicator approach, impact of macro-economic effects).
- Model benchmarking (internal versus external benchmarking, benchmarking statistics).
- Qualitative validation of analytical models (data quality and master data management, model design, documentation, involvement of management and corporate governance)
- Case study: backtesting a customer scoring model.
- Semi-supervised learning.
- Fuzzy techniques.
- Evolutionary algorithms.
- Ant colony optimization.
- Online analytics applications.
- Social media analytics applications.
- Process analytics applications.