Credit scoring by feature-weighted support vector machines

Shi, Jian; Zhang, Shuyou; Qiu, Lemiao

doi:10.1631/jzus.c1200205

Cited by 21 publications

(13 citation statements)

References 18 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In order to improve this accuracy, some studies have been made toward enhancing SVM models. Shi, et al [15] proposed a novel feature-weighted support vector machine (SVM) credit scoring model, in which an F-score is adopted for feature im-portance ranking. Focused on the performance of SVM models depends on their parameters' setting, Zhou, et al [16] used direct search method to optimize the SVM-based credit scoring model and compared it with other three parameters optimization methods, such as grid search, method based on design of experiment (DOE) and genetic algorithm (GA).…”

Section: Aim and Contributions Of This Studymentioning

confidence: 99%

Investigation of the Impact of Data Comparability on Performance of Support Vector Machine Models for Credit Scoring

Dong

Hao

Sato

2015

iscm

View full text Add to dashboard Cite

Although most of the studies of support vector machines (SVM) models focused on the algorithm improvement or parameters tuning, the performance of SVM models also depends on datasets, based on which the models were constructed. This paper investigate the impact of data comparability on performance of SVM models for credit scoring. After giving several examinations into data comparability and its impairing factors, we collect two practical datasets for credit scoring and then carry out several experiments to construct and test SVM models. According to the experiments' results, it has been clarified that SVM models can classify training datasets perfectly whatever data comparability may be, if we choose appropriate kernel function and related parameters. However, the performance of SVM models to classify new data depends heavily on data comparability. If data comparability is low, the accuracy for classifying test datasets is proportionally low and fluctuates irregularly. It is obvious that guaranteeing data comparability is more important and effective than improving algorithm or turning parameters of SVM models.

show abstract

Section: Aim and Contributions Of This Studymentioning

confidence: 99%

Investigation of the Impact of Data Comparability on Performance of Support Vector Machine Models for Credit Scoring

Dong

Hao

Sato

2015

iscm

View full text Add to dashboard Cite

show abstract

“…Support Vector Machines [6] Artificial Neural Networks [7,8] Feature Selection Based using ANN [9] Ensemble ANN [10] ANN and Decision Tables [11] Evolutionary Product-ANN [12] Fuzzy Immune Learning [13] Genetic Programming [14] Genetic Programming and SVM [15] Wavelet Networks and Particle Swarm Optimization [16] Various AI Techniques [17,18] ML homogenous and hybrid approaches show promising results; nevertheless they do not overperform simpler approaches significantly. Per contra, simpler approaches ask for attention with an opportunity for efficient prediction performance.…”

Section: Table 1 ML Approaches For Credit Risk Predictionmentioning

confidence: 99%

Business Analytics using Random Forest Trees for Credit Risk Prediction: A Comparison Study

Ghatasheh

2014

IJAST

View full text Add to dashboard Cite

show abstract

“…The non-negative variable i ξ is introduced for cases where data points are not rigidly separable, typical of practical applications, such that some slack is allowed for the wrong classified input samples [4]. The regularized constant C controls the trade-off between training error and classification margin.…”

Section: A Support Vector Machinementioning

confidence: 99%

“…The highest importance estimated features tended to be normal both in the two datasets. The figure shows that some features (e.g., 1,2,3,4,5,9,11,16,21) are relatively important in German dataset, in which the F-score of the first feature reach 0.2 and features (e.g., 5,6,7,8,9) are decisive factors for the classification in Australian dataset, in which the F-score of the eighth feature is obviously larger than other features. So it is possible to reduce dimension by deleting the features with small F-score (improved F-score) one by one, and get an optimal classifying model at last.…”

Section: B Comparisons Of Different Svmsmentioning

confidence: 99%

Credit scoring with F-score based on support vector machine

Chen

Shi

2013

Proceedings 2013 International Conference on Mechatronic Sciences, Electric Engineering and Computer (MEC)

View full text Add to dashboard Cite

Credit risk management is one of the most important issues in financial research. Reliable credit scoring models are crucial for financial agencies to evaluate credit applications. In this article, a novel feature-weighted support vector machine credit scoring models are presented for credit risk assessment, in which F-score and improved F-score is adopted for feature importance calculating. These featureweighted versions of Support Vector Machine are tested against the traditional feature selection Support Vector Machine on two real-world datasets and the research results reveal the validity of the proposed method. The feature-weighted methods have optimized performance, which improved the accuracy and reduced the modeling time consumption.

show abstract

Credit scoring by feature-weighted support vector machines

Cited by 21 publications

References 18 publications

<b>Investigation of the Impact of Data Comparability on Performance of </b><b>Support Vector Machine Models for Credit Scoring </b>

<b>Investigation of the Impact of Data Comparability on Performance of </b><b>Support Vector Machine Models for Credit Scoring </b>

Business Analytics using Random Forest Trees for Credit Risk Prediction: A Comparison Study

Credit scoring with F-score based on support vector machine

Contact Info

Product

Resources

About