International Scientific Publications
© 2018 Science Events Ltd
Terms of Use  ·  Privacy Policy
Choose language English French Russian Bulgarian

Economy & Business, Volume 12, 2018

VALIDATION PROCESS FOR SCORING AND RATING MODELS USING NEURAL NETWORKS
Anatoliy Antonov, Ventsislav Nikolov
Pages: 95-104
Published: 12 Sep 2018
Views: 32
Abstract: This research paper investigates the validation and calibration of models for determination of credit scoring and rating with statistical methods. This is done through a comparison of the results of the model to an alternative model, based on a neural network, and a calculation of different statistical parameters. A prototype of a software system for analysis and evaluations is represented that calculates distance, standard deviation, correlation, cumulative accuracy profile (CAP), as well as accumulation and analysis of historical statistics for default losses.
Keywords: credit rating, scoring, validation, analysis, calibration, neural networks
Download full text

Back to the contents of the volume
By using this site you agree to our Privacy Policy and Terms of Use. We use cookies, including for analytics, personalisation, and ads.