College Admissions: Predicting Applicant Fit, Quality, and Yield to College Admissions based on Diverse Applicant Factors.

Business Context

A leading admissions software company wanted to upgrade its offering by helping colleges auto-select the right fit and high yield candidates

Key Questions

  • What are the factors that drive the quality of an applicant?
  • What are the factors that drive yield?
  • How can data-driven predictive solution be designed to select the candidates?

Impact Created

  • Created revenue potential of over $7M annually for the admissions software provider
  • Ability to predict high-quality applicant pool with 80% accuracy
  • Ability to predict high yield applicants with 70& accuracy

Solution Design

  • Variable identification and selection
  • Solutioning
  • Output

Solution Design:

Variable identification and selection:

  1. Applicant Demography
  2. Work experience
  3. Academics
  4. External Factors
  5. Application

Solutioning: Solutions Traits

  • Level Of Solution: At college
  • Time period: One year
  • Technique: AI and ML: Logistic Regression, SVM, Random forest
  • Validation: Out of Sample and Out of time

Results:

Tags: Education | Marketing