Model Development and Validation

Description

COURSE DETAILS:

 

Day 1: Model Development

Description

This training dives deep into the statistical and analytical aspects of credit risk modeling. It covers advanced modeling techniques, the complete model development and validation process, and the application of machine learning and AI in risk assessment.

 

Target Audience:

Risk management professionals, data scientists, risk modelers, quantitative analysts, and other technical professionals

 

Course Outline:

  1. Introduction to Credit Risk Scoring
    • Basics: Credit, Credit Risk, Credit Score
    • Role of credit scoring in credit risk management
    • BSP Circular 855 requirements for model development
    • Types of scorecards (application, behavioral, collection) and their use cases

 

  1. Model Development Lifecycle
    • Stages: Business requirements → Data collection → Data preparation → Modeling → Implementation → Monitoring
    • Key stakeholders and governance in model development

 

  1. Defining Project Parameters
    • Business objectives and scope definition
    • Portfolio segmentation strategies
    • Observation and performance window selection
    • “Good-Bad-Indeterminate-Exclusions” definitions

 

  1. Data Preparation
    • Data sources: application, behavioral, credit bureau
    • Data quality checks, cleansing, and transformation
    • Handling missing values, outliers, and derived variables

 

  1. Variable Selection & Transformation
    • Initial characteristic analysis and Information Value (IV) computation
    • Bivariate analysis and business validation
    • Data transformations (binning, capping, WOE conversion)

 

  1. Model Building Techniques
    • Logistic regression fundamentals
    • Stepwise selection methods (forward, backward, hybrid)
    • Reject inference techniques
    • Scaling and scorecard points assignment
    • Machine learning techniques: Gradient boosting, random forest, explainability

 

  1. Documentation and Handover
    • Regulatory documentation requirements
    • Model user guide and technical specification

 

  1. Additional Topics
    • Alternative Data in Credit Scoring
    • Machine Learning in Credit Risk Models – Gradient boosting, random forests,
    • IFRS 9 Alignment – Using PD, LGD, EAD models for
    • Ethical AI and model fairness in
    • Decision Engine Integration
    • Portfolio Simulation and What-if Analysis

 

Day 2: Model Validation

Description

This course provides a comprehensive overview of model validation in the context of risk management, compliance and audit. It focuses on understanding the fundamental concepts, regulatory requirements and bestpractices. The course aims to equip participants with the necessary skills to effectively conduct model validation.It covers qualitative and quantitative validation techniques, including industry-specific case studies.

 

Target Audience:

Risk management professionals, internal auditors, regulatory compliance personnel

 

Course Outline

  1. Introduction to Model Validation
    • Regulatory requirements (BSP Circular 855, IFRS 9)
    • Model risk management principles

 

  1. Validation Framework
    • Independent validation development team role separation
    • Types of validation: developmental evidence, outcome analysis, process validation

 

  1. Quantitative Validation 
    • Discriminatory power: KS, Gini, AUC, divergence measures
    • Calibration: accuracy ratio, Hosmer–Lemeshow test
    • Stability metrics: PSI, characteristic stability index

 

  1. Qualitative Validation
    • Review of development methodology, data sources, and assumptions
    • Alignment with business processes and policies
    • Stress testing and sensitivity analysis

 

  1. Back-testing and Benchmarking
    • Out-of-time sample validation
    • Comparing model results with challenger models or external benchmarks

 

  1. Monitoring and Governance
    • Ongoing model monitoring and triggers for redevelopment
    • Score-to-odds maintenance and cutoff review
    • Reporting requirements for management and regulators

 

  1. Remediation and Model Redevelopment
    • Handling performance deterioration
    • Updating variables and recalibration
    • Version control and audit trail

 

  1. Additional Topics
    • Vintage analysis – for tracking performance drift over time
    • Roll Rates analysis – for validating bad definitions
    • Strategy validation
    • Segment-level drift analysis
    • Portfolio impact simulations
    • Champion-Challenger monitoring
    • Model monitoring framework
    • Triggers for recalibration and redevelopment

 

RESOURCE SPEAKER:

Mr. Ricardo Ian P. Copia
Training Consultant

 

SCHEDULE:

October 28-29, 2025 (Tuesday-Wednesday) 9:00 AM – 5:00 PM

 

Training Fee Per Participant:

Member Institution – Php7,840.00

Non-Member Institution – Php10,080.00

*VAT inclusive

 

REGISTER HERE!