Finite-Difference Approximation for Explaining Neural Network Predictions in Financial Credit Evaluation

Overview

Uses finite-difference approximations to characterize how input changes affect neural-network credit predictions.

Research context

This work contributes to Interpretability. The publication record above is the authoritative source for its current status; accepted papers will be updated when final bibliographic metadata becomes public.

@inproceedings{FiniteDifferenceCreditXai2025,
  title = {Finite-Difference Approximation for Explaining Neural Network Predictions in Financial Credit Evaluation},
  author = {M. Aziz, M. U. Danish, F. Iqbal,, U. Rehman},
  year = {2025},
  booktitle = {International Conference on Advanced Machine Learning and Data Science}
}