Primary tabs

Short Course
on
Building predictive models using SAS Enterprise Miner: Credit scoring case study
The financial industry currently has a shortage of people with the necessary data science skills to address problems relating to large and/or complex data sets. Nowadays financial institutions are confronted with an abundance of data, and so in order to maintain a competitive advantage it is imperative to have analysts with these necessary data science skills. This short course will address the topic of Building predictive models.

Purpose of the course:

This short course aims to equip participants with the skills necessary to build predictive models in SAS Enterprise Miner (using a credit scorecard as case study).

Admission requirements:

Admission requirements: 
NQF level 7 majoring in either Mathematics/Statistics/Engineering/Computer Science.
Learning assumed to be in place: 
NQF level 7 (previous knowledge on credit scoring or predictive modelling is recommended)

Course outcomes and assessment criteria :

Course outcomes and the associated assessment criteria: 

Study Unit

Outcomes

Assessment Criteria

  After completion of this course, participants will: Participant will be assessed on the following criteria:
 
  1. Demonstrate applied knowledge and understanding of building predictive models in SAS Enterprise Miner and understand how it is applied in practice.
  2. Demonstrate an ability to identify real-world practical problems that can be solved using predictive models and then apply statistical methods to solve these problems.
  3. Demonstrate an ability to use a range of predictive modelling skills to identify, analyse and address complex problems
  4. Demonstrate an understanding of the preliminaries and planning involved in building a scorecard as well as an understanding of the data requirements and project parameters for building a credit scorecard
  1. Critically discuss and execute an exploratory analysis for predictive modelling using SAS EM on a dataset;
  2. Implement appropriate procedures to select the appropriate variables on a dataset using SAS EM.
  3. Implement appropriate procedures to bin the variables on a simplified dataset and evaluate the quality of the chosen binning, employing weights of evidence, information value and logical trend using SAS EM;
  4. Develop a basic logistic regression model in SAS EM
  5. Implement appropriate procedures to build a preliminary predictive model (i.e. scorecard) in SAS EM
  6. Implement appropriate procedures to build alternative predictive models in SAS EM using Decision Trees
  7. Critically evaluate the different predictive models built in SAS EM
  8. Perform reject inference on a dataset, as well as to develop a final predictive model on this dataset and validate the scorecard built (in SAS EM)
  9. Apply predictive modelling principles during the discussion of a  case study on a real-world credit scoring dataset

 

Assessment: 
Assignments and participation
Method of assessment: 
Learning objectives will be accomplished through the successful completion of assignments and participation in short course activities, proving insight into the topics at hand.

Additional information

Mode of delivery: 
Mixed
Target group: 
Graduates and/or industry professionals working in fields that require data science and analytical skills.
Contact us
Name: 
Prof Tanja Verster
Telephone number: 
018 299 2566