haku: @author Tessmer, A. C. / yhteensä: 2
viite: 1 / 2
« edellinen | seuraava »
Tekijä: | Tessmer, A. C. |
Otsikko: | What to learn from near misses: An inductive learning approach to credit risk assessment |
Lehti: | Decision Sciences
1997 : WINTER, VOL. 28:1, p. 105-120 |
Asiasana: | FINANCE DECISION SUPPORT SYSTEMS DECISION THEORY CREDIT FINANCIAL RISK |
Kieli: | eng |
Tiivistelmä: | A new dimension of inductive learning for credit risk analysis is presented in this article, based on the specific impact of Type I and Type II credit errors on the accuracy of the learning process. The near misses can nudge the learning process towards a more accurate definition of the boundary between negative and positive examples. The results support the assumptions by showing that the learning process is more accurate and stable after the Dynamic Updating Process is performed, and they indicate that an inductive learning approach ignoring the presence of the near missed examples and their potential contribution to a more accurate concept remains incomplete. Whenever a Type I/II error is detected, the Dynamic Updating Process should actually be performed. |
« edellinen | seuraava »
SCIMA