Decision Matrix

Decision Matrix

The Matrix Decision model in Scorport enables users to create decision frameworks by analyzing the intersection of two independent scorecards. This tool is commonly used in credit scoring scenarios to visually categorize risks and define decisions for each intersection dynamically.


What Is a Matrix Decision?

Matrix Decisions allow you to cross-analyze two scorecards by defining vertical and horizontal axes with dynamic categories. Each category represents a risk level, and users can assign specific decisions (e.g., Yes, Maybe, No) for every combination of risk levels in the grid. Scorport offers complete flexibility to customize the number of categories and their boundaries.


Key Features

  1. Two Independent Inputs:

    • Define the vertical axis and horizontal axis using two separate scorecards.
    • Example:
      • Vertical Axis: Credit Bureau Score (Behavior Score).
      • Horizontal Axis: Application Score.
  2. Predefined Risk Groups:

    • The axes are divided into risk categories:
      • Excellent (A), Good (B), Middle (C), Poor (D), Bad (E).
    • These risk groups are commonly used in credit bureau scoring systems for simplicity and clarity.
  3. Dynamic Grid Customization:

    • Users can dynamically:
      • Add or remove categories (e.g., from 3 to 5 or more) by clicking Add.
      • Adjust the score ranges for each category.
  4. Real-Time Status Updates:

    • Assign statuses (e.g., Yes, No, Maybe) and colors to each cell with a single click.
    • Adjust the decision logic dynamically without reconfiguring the entire grid.

How to Create a Matrix Decision

Step 1: Navigate to the Decision Designer

  • Open the Decision Designer module from the dashboard.
  • Select Matrix Decision as your decision model type.

Step 2: Define Axes and Categories

  1. Vertical Axis:

    • Assign the first scorecard (e.g., Credit Bureau Score) to the vertical axis.
    • Divide the axis into predefined risk groups:
      • Excellent (A): Top-tier score range.
      • Good (B): Above-average score range.
      • Middle (C): Average score range.
      • Poor (D): Below-average score range.
      • Bad (E): Lowest score range.
    • Customize the boundaries for each risk group using the (From, To] convention.
  2. Horizontal Axis:

    • Assign the second scorecard (e.g., Application Score) to the horizontal axis.
    • Use the same risk groups as the vertical axis or define a custom set of categories.
  3. Dynamic Customization:

    • Click Add to increase the number of categories on either axis.
    • Adjust the score ranges for each category to match your business logic.

Step 3: Configure Cell Decisions

  • Click on any cell in the matrix to assign:
    • Label: Represents the combination (e.g., BB, DA).
    • Status: Assign a decision status (e.g., Yes, Maybe, No).
    • Color: Use colors to visually represent outcomes:
      • Green: Positive outcomes (e.g., Yes).
      • Yellow: Conditional outcomes (e.g., Maybe).
      • Red: Negative outcomes (e.g., No).

Example: Credit Risk Analysis

Scenario:

  • Vertical Axis: Credit Bureau Score (Behavior Score).

    • Excellent (A): (900, 1000]
    • Good (B): (750, 900]
    • Middle (C): (500, 750]
    • Poor (D): (300, 500]
    • Bad (E): (0, 300]
  • Horizontal Axis: Application Score.

    • Excellent (A): (800, 1000]
    • Good (B): (650, 800]
    • Middle (C): (400, 650]
    • Poor (D): (200, 400]
    • Bad (E): (0, 200]

image.png

Outcomes:

  • Cell AA (High Behavior + High Application): Yes (Green) → Approve Premium Loan.
  • Cell CC (Middle Behavior + Middle Application): Maybe (Yellow) → Further Review.
  • Cell ED (Bad Behavior + Poor Application): No (Red) → Deny Loan.

Benefits of Matrix Decisions

  • Customizable Risk Categories: Add or adjust categories dynamically to suit your scoring model.
  • Granular Decision Control: Define specific outcomes for every scorecard combination.
  • Real-Time Adjustments: Easily update decisions and boundaries to reflect changing business needs.

Matrix Decisions are ideal for cross-analyzing scorecards and implementing granular decision-making logic.