Multiple classifier architectures and their application to credit risk assessment

Finlay, Steven

doi:10.1016/j.ejor.2010.09.029

Cited by 185 publications

(104 citation statements)

References 49 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…To illustrate the standardization process, consider "Age" in Row 71 of Table 3. The age of the legal person associated with loan X2012060800099 in Column 1231 is 38, which is within the optimum interval [31]. According to Equation (3c), this makes the standardized value x ij of "Age" in this case equal to 1.…”

Section: Establishing the Credit Rating Index Systemmentioning

confidence: 99%

“…Florez-Lopez proposed a variety of statistical methods (multiple discriminant analysis, multinomial logit regression, and ordered logit) and methods associated with decision trees to explore the determinants of ratings [30]. Finlay presented a multi-classifier system and introduced a new boosting algorithm, dubbed ET Boost [31]. Finally, Paleologo et al created a method for adding missing data and proposed the use of an ensemble classification technique called subagging [32].…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Multi Criteria Credit Rating Model for Small Enterprise Using a Nonparametric Method

Chi

Zhang

2017

Sustainability

View full text Add to dashboard Cite

Abstract:A small enterprise's credit rating is employed to measure its probability of defaulting on a debt, but, for small enterprises, financial data are insufficient or even unreliable. Thus, building a multi criteria credit rating model based on the qualitative and quantitative criteria is of importance to finance small enterprises' activities. Till now, there has not been a multicriteria credit risk model based on the rank sum test and entropy weighting method. In this paper, we try to fill this gap by offering three innovative contributions. First, the rank sum test shows significant differences in the average ranks associated with index data for the default and entire sample, ensuring that an index makes an effective differentiation between the default and non-default sample. Second, the rating equation's capacity is tested to identify the potential defaults by verifying a clear difference between the average ranks of samples with default ratings (i.e., not index values) and the entire sample. Third, in our nonparametric test, the rank sum test is used with rank correlation analysis made to screen for indices, thereby avoiding the assumption of normality associated with more common credit rating methods.

show abstract

Section: Establishing the Credit Rating Index Systemmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Multi Criteria Credit Rating Model for Small Enterprise Using a Nonparametric Method

Chi

Zhang

2017

Sustainability

View full text Add to dashboard Cite

show abstract

“…As noted, automated lending risk evaluations are currently imperfect and, in fact, the failure of credit scoring algorithms to identify loan recipients who will eventually default on their loans results in sizable losses on an ongoing basis [3].…”

Section: Introductionmentioning

confidence: 99%

Recursive-Rule Extraction Algorithm With J48graft And Applications To Generating Credit Scores

Hayashi

Tanaka

Takagi

et al. 2016

Journal of Artificial Intelligence and Soft Computing Research

View full text Add to dashboard Cite

The purpose of this study was to generate more concise rule extraction from the RecursiveRule Extraction (Re-RX) algorithm by replacing the C4.5 program currently employed in Re-RX with the J48graft algorithm. Experiments were subsequently conducted to determine rules for six different two-class mixed datasets having discrete and continuous attributes and to compare the resulting accuracy, comprehensibility and conciseness. When working with the CARD1, CARD2, CARD3, German, Bene1 and Bene2 datasets, Re-RX with J48graft provided more concise rules than the original Re-RX algorithm. The use of Re-RX with J48graft resulted in 43.2%, 37% and 21% reductions in rules in the case of the German, Bene1 and Bene2 datasets compared to Re-RX. Furthermore, the Re-RX with J48graft showed 8.87% better accuracy than the Re-RX algorithm for the German dataset. These results confirm that the application of Re-RX in conjunction with J48graft has the capacity to facilitate migration from existing data systems toward new concise analytic systems and Big Data.

show abstract

“…The linear combination models proposed by Shen, Li and Song (2008, 2011) and Wong et al (2007 are used to examine the optimal subset selection algorithm for this study. The information concepts have never been applied to the selection of individual models as combination models, and all available individual models are used as inputs for the linear combination methods in tourism demand forecasting literature.…”

Section: Introductionmentioning

confidence: 99%

“…Abellán and Masegosa (2010) proposed the ensemble method using credal decision trees, and showed the good percentage of correct classifications and an improvement in time of processing, especially for large data sets. Finlay (2011) applied bagging and boosting methods to the credit risk assessment to classify consumers as good or bad credit risks, and proposed a new boosting algorithm, 'error trimmed boosting'. Experiments showed that the bagging and boosting methods outperform other multi-classifier systems, and 'error trimmed boosting' outperforms bagging and AdaBoost by a significant margin.…”

Section: Introductionmentioning

confidence: 99%

A combination selection algorithm on forecasting

Cang

2014

European Journal of Operational Research

View full text Add to dashboard Cite

It is widely accepted in forecasting that a combination model can improve forecasting accuracy. One important challenge is how to select the optimal subset of individual models from all available models without having to try all possible combinations of these models. This paper proposes an optimal subset selection algorithm from all individual models using information theory. The experimental results in tourism demand forecasting demonstrate that the combination of the individual models from the selected optimal subset significantly outperforms the combination of all available individual models. The proposed optimal subset selection algorithm provides a theoretical approach rather than experimental assessments which dominate literature.

show abstract

Multiple classifier architectures and their application to credit risk assessment

Cited by 185 publications

References 49 publications

Multi Criteria Credit Rating Model for Small Enterprise Using a Nonparametric Method

Multi Criteria Credit Rating Model for Small Enterprise Using a Nonparametric Method

Recursive-Rule Extraction Algorithm With J48graft And Applications To Generating Credit Scores

A combination selection algorithm on forecasting

Contact Info

Product

Resources

About