Center for Machine Learning and Intelligent Systems
About  Citation Policy  Donate a Data Set  Contact


Repository Web            Google
View ALL Data Sets

Credit Approval Data Set

Below are papers that cite this data set, with context shown. Papers were automatically harvested and associated with this data set, in collaboration with Rexa.info.

Return to Credit Approval data set page.


Xiaoming Huo. FBP: A Frontier-Based Tree-Pruning Algorithm. Seoung Bum Kim. 2002.

the mean difference of the CV error between CCP and FBP is (0.0770, 0.2196). As mentioned earlier, we treat this as a "sanity check". Table 4: Comparison of the CV Error Rates Between CCP and FBP Data Set CCP FBP Winner Australian Credit Approval 14.13 14.01 FBP Cleveland Heart Disease 21.15 20.89 FBP Congressional Voting Records 4.16 4.12 FBP Wisconsin Breast Cancer 4.56 4.47 FBP Iris Plants 5.20


Lorne Mason and Peter L. Bartlett and Jonathan Baxter. Improved Generalization Through Explicit Optimization of Margins. Machine Learning, 38. 2000.

chosen as the final solution. In some cases the training sets were reduced in size to makeoverfitting more likely (so that complexity regularization with DOOM could have an effect). In three of the datasets Credit Application Wisconsin Breast Cancer and Pima Indians Diabetes), AdaBoost gained no advantage from using more than a single classifier. In these datasets, the number of classifiers was


Kagan Tumer and Joydeep Ghosh. Robust Combining of Disparate Classifiers through Order Statistics. CoRR, csLG/9905013. 1999.

and the corresponding size of the MLP used, are 5 : ffl Card: a 51-dimensional, 2-class data set based on credit approval decision with 690 patterns; an MLP with 10 hidden units; ffl Gene: a 120-dimensional data set with two classes, based on the detection of splice junctions in DNA sequences,


Lorne Mason and Peter L. Bartlett and Jonathan Baxter. Direct Optimization of Margins Improves Generalization in Combined Classifiers. NIPS. 1998.

sets were reduced in size to makeoverfitting more likely, so that complexity regularization with DOOM could haveaneffect. (The details are given in the full version [MBB98].) In three of the datasets Credit Application Wisconsin Breast Cancer and Pima Indians Diabetes), AdaBoost gained no advantage from using more than a single classifier. In these datasets, the number of classifiers was


Return to Credit Approval data set page.

Supported By:

 In Collaboration With:

About  ||  Citation Policy  ||  Donation Policy  ||  Contact  ||  CML