Advanced level · Module 2: Predictive Modelling in Data Science Projects · Lesson 3 of 4
Inclusion and Exclusion Errors
Evaluate targeting models by who is wrongly included and wrongly excluded, overall and by area.
In this lesson you will
- Calculate inclusion and exclusion error rates
- Choose cut-offs considering budget and fairness
Key concepts
inclusion errorexclusion errorcoverage and budgetgroup-wise errors
The lesson
Targeting errors have names: inclusion errors give assistance to households not in need; exclusion errors leave needy households out. Exclusion errors usually do greater harm.
The cut-off is set by budget and priorities: covering 30% of households gives a different balance of errors than covering 50%.
Calculate errors by area, gender of household head and other groups. A model that works well overall may exclude many needy households in remote areas.
Step by step
- Calculate inclusion and exclusion errors at two coverage levels.
- Break exclusion errors down by area.
- Compare with the current rule.
- Recommend a cut-off.
Practice activity
Complete the targeting error analysis from the practical activity.
Expected result: Error tables by group with a justified recommendation.
Check your understanding
-
What is an exclusion error in targeting?
Show answer
A household in need that is not selected.
-
Why break errors down by area?
Show answer
Good overall performance can hide high exclusion in some areas.
Taught live by an instructor — online or in person. Free course lessons are for self-study and do not lead to a certificate — see registration options.