Understading black boxes: Knowledge induction from models

Jinhwa Kim, Jaekwon Bae, Dursun Delen

Research output: Contribution to conferencePaper

Abstract

Due to regurations and laws prohibiting uses of private data on customers and their transactions in customer data base, most customer data sets are not easily accessable even in the same organizations. A solutio for this reguatory problems can be providing statistical summary of the data or models induced from the dat, instead of providing raw data sets. The models, however, have limited information on the original raw data set. This study explores possible solutions for these problems. The study uses prediction models from data on credit information of customers provided by a local bank in Seoul, S. Korea. This study suggests approaches in figuring what is inside of the non-rules based models such as regression models or neural network models. The study proposes several rule accumulation algorithms such as (RAA) and a GA-based rule refinement algorithm (GA-RRA) as possible solutions for the problems. The experiments show the performance of the random dataset, RAA, elimination of redundant rules (ERR), and GA-RRA.

Original languageEnglish
StatePublished - 1 Dec 2011
Externally publishedYes
Event15th Pacific Asia Conference on Information Systems: Quality Research in Pacific, PACIS 2011 - Brisbane, QLD, Australia
Duration: 7 Jul 201111 Jul 2011

Conference

Conference15th Pacific Asia Conference on Information Systems: Quality Research in Pacific, PACIS 2011
CountryAustralia
CityBrisbane, QLD
Period7/07/1111/07/11

Fingerprint

Neural networks
Experiments

Keywords

  • Ga-based rule refinement algorithm (GA-RRA)
  • Logistic regression model
  • Personal credit rating
  • Rule accumulation algorithm (RAA)

Cite this

Kim, J., Bae, J., & Delen, D. (2011). Understading black boxes: Knowledge induction from models. Paper presented at 15th Pacific Asia Conference on Information Systems: Quality Research in Pacific, PACIS 2011, Brisbane, QLD, Australia.
Kim, Jinhwa ; Bae, Jaekwon ; Delen, Dursun. / Understading black boxes : Knowledge induction from models. Paper presented at 15th Pacific Asia Conference on Information Systems: Quality Research in Pacific, PACIS 2011, Brisbane, QLD, Australia.
@conference{25af292321054d248d129f69e2ca48b8,
title = "Understading black boxes: Knowledge induction from models",
abstract = "Due to regurations and laws prohibiting uses of private data on customers and their transactions in customer data base, most customer data sets are not easily accessable even in the same organizations. A solutio for this reguatory problems can be providing statistical summary of the data or models induced from the dat, instead of providing raw data sets. The models, however, have limited information on the original raw data set. This study explores possible solutions for these problems. The study uses prediction models from data on credit information of customers provided by a local bank in Seoul, S. Korea. This study suggests approaches in figuring what is inside of the non-rules based models such as regression models or neural network models. The study proposes several rule accumulation algorithms such as (RAA) and a GA-based rule refinement algorithm (GA-RRA) as possible solutions for the problems. The experiments show the performance of the random dataset, RAA, elimination of redundant rules (ERR), and GA-RRA.",
keywords = "Ga-based rule refinement algorithm (GA-RRA), Logistic regression model, Personal credit rating, Rule accumulation algorithm (RAA)",
author = "Jinhwa Kim and Jaekwon Bae and Dursun Delen",
year = "2011",
month = "12",
day = "1",
language = "English",
note = "null ; Conference date: 07-07-2011 Through 11-07-2011",

}

Kim, J, Bae, J & Delen, D 2011, 'Understading black boxes: Knowledge induction from models' Paper presented at 15th Pacific Asia Conference on Information Systems: Quality Research in Pacific, PACIS 2011, Brisbane, QLD, Australia, 7/07/11 - 11/07/11, .

Understading black boxes : Knowledge induction from models. / Kim, Jinhwa; Bae, Jaekwon; Delen, Dursun.

2011. Paper presented at 15th Pacific Asia Conference on Information Systems: Quality Research in Pacific, PACIS 2011, Brisbane, QLD, Australia.

Research output: Contribution to conferencePaper

TY - CONF

T1 - Understading black boxes

T2 - Knowledge induction from models

AU - Kim, Jinhwa

AU - Bae, Jaekwon

AU - Delen, Dursun

PY - 2011/12/1

Y1 - 2011/12/1

N2 - Due to regurations and laws prohibiting uses of private data on customers and their transactions in customer data base, most customer data sets are not easily accessable even in the same organizations. A solutio for this reguatory problems can be providing statistical summary of the data or models induced from the dat, instead of providing raw data sets. The models, however, have limited information on the original raw data set. This study explores possible solutions for these problems. The study uses prediction models from data on credit information of customers provided by a local bank in Seoul, S. Korea. This study suggests approaches in figuring what is inside of the non-rules based models such as regression models or neural network models. The study proposes several rule accumulation algorithms such as (RAA) and a GA-based rule refinement algorithm (GA-RRA) as possible solutions for the problems. The experiments show the performance of the random dataset, RAA, elimination of redundant rules (ERR), and GA-RRA.

AB - Due to regurations and laws prohibiting uses of private data on customers and their transactions in customer data base, most customer data sets are not easily accessable even in the same organizations. A solutio for this reguatory problems can be providing statistical summary of the data or models induced from the dat, instead of providing raw data sets. The models, however, have limited information on the original raw data set. This study explores possible solutions for these problems. The study uses prediction models from data on credit information of customers provided by a local bank in Seoul, S. Korea. This study suggests approaches in figuring what is inside of the non-rules based models such as regression models or neural network models. The study proposes several rule accumulation algorithms such as (RAA) and a GA-based rule refinement algorithm (GA-RRA) as possible solutions for the problems. The experiments show the performance of the random dataset, RAA, elimination of redundant rules (ERR), and GA-RRA.

KW - Ga-based rule refinement algorithm (GA-RRA)

KW - Logistic regression model

KW - Personal credit rating

KW - Rule accumulation algorithm (RAA)

UR - http://www.scopus.com/inward/record.url?scp=84862922913&partnerID=8YFLogxK

M3 - Paper

AN - SCOPUS:84862922913

ER -

Kim J, Bae J, Delen D. Understading black boxes: Knowledge induction from models. 2011. Paper presented at 15th Pacific Asia Conference on Information Systems: Quality Research in Pacific, PACIS 2011, Brisbane, QLD, Australia.