Credit Scoring And Its Applications By L C Thomas Hot ^hot^ Here

A recurring theme in Thomas’s work is rejection inference : how do you validate a model when you only observe outcomes for approved applicants? He championed and expectation-maximization methods long before they became machine learning staples.

Lenders are now using LLMs (Large Language Models) to generate synthetic borrower histories to train models where real data is scarce (e.g., pandemic-era defaults). credit scoring and its applications by l c thomas hot