About this document
Stochastic Feature Selection for Credit Segmentation by akbar is a document available to read on EtoBox.
The document proposes combining k-means clustering with stochastic local search for feature selection to segment clients in credit scoring. It applies this technique to five financial datasets to cluster clients as "good" or "bad" credit risks. The results show benefits for banks and customer segmentation by minimizing within-cluster variance compared to standard k-means.
- Author
- akbar
- Language
- EN