Balancing accuracy, complexity and interpretability in consumer credit decision making: A C-TOPSIS classification approach

Published on Nov 1, 2013in Knowledge Based Systems5.101
· DOI :10.1016/j.knosys.2013.08.004
Xiaoqian Zhu7
Estimated H-index: 7
(CAS: Chinese Academy of Sciences),
Jianping LiXiaolei20
Estimated H-index: 20
(CAS: Chinese Academy of Sciences)
+ 2 AuthorsChangzhi Liang5
Estimated H-index: 5
(CAS: Chinese Academy of Sciences)
Accuracy, complexity and interpretability are very important in credit classification. However, most approaches cannot perform well in all the three aspects simultaneously. The objective of this study is to put forward a classification approach named C-TOPSIS that can balance the three aspects well. C-TOPSIS is based on the rationale of TOPSIS (Technique for Order Preference by Similarity to Ideal Solution). TOPSIS is famous for reliable evaluation results and quick computing process and it is easy to understand and use. However, it is a ranking approach and three challenges have to be faced for modifying TOPSIS into a classification approach. C-TOPSIS works out three strategies to overcome the challenges and retains the advantages of TOPSIS. So C-TOPSIS is deduced to have reliable classification results, high computational efficiency and ease of use and understanding. Our findings in the experiment verify the advantages of C-TOPSIS. In comparison with 7 popular approaches on 2 widely used UCI credit datasets, C-TOPSIS ranks 2nd in accuracy, 1st in complexity and is in 1st rank in interpretability. Only C-TOPSIS ranks among the top 3 in all the three aspects, which verifies that C-TOPSIS can balance accuracy, complexity and interpretability well.
  • References (44)
  • Citations (23)
📖 Papers frequently viewed together
10 Citations
6,178 Citations
101 Citations
78% of Scinapse members use related papers. After signing in, all features are FREE.
#1Lu Han (THU: Tsinghua University)H-Index: 2
#2Liyan Han (Beihang University)H-Index: 2
Last. Hongwei Zhao (THU: Tsinghua University)H-Index: 2
view all 3 authors...
The most commonly used techniques for credit scoring is logistic regression, and more recent research has proposed that the support vector machine is a more effective method. However, both logistic regression and support vector machine suffers from curse of dimension. In this paper, we introduce a new way to address this problem which is defined as orthogonal dimension reduction. We discuss the related properties of this method in detail and test it against other common statistical approaches-pr...
22 CitationsSource
#1You-Shyang ChenH-Index: 12
#2Ching-Hsue Cheng (National Yunlin University of Science and Technology)H-Index: 37
Banks are important to national, and even global, economic stability. Banking panics that follow bank insolvency or bankruptcy, especially of large banks, can severely jeopardize economic stability. Therefore, issuers and investors urgently need a credit rating indicator to help identify the financial status and operational competence of banks. A credit rating provides financial entities with an assessment of credit worthiness, investment risk, and default probability. Although numerous models h...
51 CitationsSource
#1Majid BehzadianH-Index: 10
#2S. Khanmohammadi Otaghsara (IAU: Islamic Azad University)H-Index: 1
Last. Joshua Ignatius (Universiti Sains Malaysia)H-Index: 16
view all 4 authors...
Multi-Criteria Decision Aid (MCDA) or Multi-Criteria Decision Making (MCDM) methods have received much attention from researchers and practitioners in evaluating, assessing and ranking alternatives across diverse industries. Among numerous MCDA/MCDM methods developed to solve real-world decision problems, the Technique for Order Preference by Similarity to Ideal Solution (TOPSIS) continues to work satisfactorily across different application areas. In this paper, we conduct a state-of-the-art lit...
592 CitationsSource
#1Ling-Jing Kao (NTUT: National Taipei University of Technology)H-Index: 1
#2Chih-Chou Chiu (NTUT: National Taipei University of Technology)H-Index: 1
Last. Fon-Yu Chiu (NTUT: National Taipei University of Technology)H-Index: 1
view all 3 authors...
A Bayesian latent variable model with classification and regression tree approach is built to overcome three challenges encountered by a bank in credit-granting process. These three challenges include (1) the bank wants to predict the future performance of an applicant accurately; (2) given current information about cardholders' credit usage and repayment behavior, financial institutions would like to determine the optimal credit limit and APR for an applicant; and (3) the bank would like to imp...
27 CitationsSource
#1Tsui-Chih Wu (Shih Chien University)H-Index: 1
#2Ming-Fu Hsu (NCNU: National Chi Nan University)H-Index: 7
The sub-prime mortgage crisis of 2007 and the global financial tsunami of 2008 have undermined worldwide economic stability. Consequently, the timely analysis of credit risk has become more essential than ever before. The performance of early risk warning mechanisms may vary according to the criteria used and the underlying environment. This study establishes numerous criteria to assess the performance of classifiers and introduces a multiple criteria decision making method to determine suitable...
32 CitationsSource
#1Ching-Chiang YehH-Index: 1
#2Fengyi Lin (NTUT: National Taipei University of Technology)H-Index: 3
Last. Chih-Yu Hsu (NTUT: National Taipei University of Technology)H-Index: 1
view all 3 authors...
In current credit ratings models, various accounting-based information are usually selected as prediction variables, based on historical information rather than the market's assessment for future. In the study, we propose credit rating prediction model using market-based information as a predictive variable. In the proposed method, Moody's KMV (KMV) is employed as a tool to evaluate the market-based information of each corporation. To verify the proposed method, using the hybrid model, which com...
45 CitationsSource
#1Jianping LiXiaolei (CAS: Chinese Academy of Sciences)H-Index: 20
#2Gang Li (CAS: Chinese Academy of Sciences)H-Index: 1
Last. Cheng-Few Lee (RU: Rutgers University)H-Index: 25
view all 4 authors...
Credit risk analysis has long attracted great attention from both academic researchers and practitioners. However, the recent global financial crisis has made the issue even more important because of the need for further enhancement of accuracy of classification of borrowers. In this study an evolution strategy (ES) based adaptive L"q SVM model with Gauss kernel (ES-AL"qG-SVM) is proposed for credit risk analysis. Support vector machine (SVM) is a classification method that has been extensively ...
11 CitationsSource
#1Sanja Vukovic (University of Belgrade)H-Index: 1
#2Boris Delibasic (University of Belgrade)H-Index: 10
Last. Milija Suknovic (University of Belgrade)H-Index: 7
view all 4 authors...
We propose a case-based reasoning (CBR) model that uses preference theory functions for similarity measurements between cases. As it is hard to select the right preference function for every feature and set the appropriate parameters, a genetic algorithm is used for choosing the right preference functions, or more precisely, for setting the parameters of each preference function, as to set attribute weights. The proposed model is compared to the well-known k-nearest neighbour (k-NN) model based ...
36 CitationsSource
#1Akhil Bandhu Hens (IIT-KGP: Indian Institute of Technology Kharagpur)H-Index: 1
#2Manoj Kumar Tiwari (IIT-KGP: Indian Institute of Technology Kharagpur)H-Index: 47
With the rapid growth of credit industry, credit scoring model has a great significance to issue a credit card to the applicant with a minimum risk. So credit scoring is very important in financial firm like bans etc. With the previous data, a model is established. From that model is decision is taken whether he will be granted for issuing loans, credit cards or he will be rejected. There are several methodologies to construct credit scoring model i.e. neural network model, statistical classific...
50 CitationsSource
#1Gang Wang (Hefei University of Technology)H-Index: 10
#2Jian Ma (CityU: City University of Hong Kong)H-Index: 39
Last. Kaiquan Xu (CityU: City University of Hong Kong)H-Index: 13
view all 4 authors...
101 CitationsSource
Cited By23
#1Tajul Rosli Razak (University of Nottingham)H-Index: 4
#2Jonathan M. Garibaldi (University of Nottingham)H-Index: 34
Last. Christian Wagner (University of Nottingham)H-Index: 18
view all 3 authors...
Hierarchical fuzzy systems (HFSs) have been seen as an effective approach to reduce the complexity of fuzzy logic systems (FLSs), largely as a result of reducing the number of rules. However, it is not clear completely how complexity of HFSs can be measured. In FLSs, complexity is commonly expressed using a multi-factorial approach, taking into consideration the number of rules, variables, and fuzzy terms. However, this may not be the best way to assess complexity in HFSs that have structures in...
#1Feng Shen (SWUFE: Southwestern University of Finance and Economics)
#2Run Wang (SWUFE: Southwestern University of Finance and Economics)
Last. Yu Shen (SWUFE: Southwestern University of Finance and Economics)
view all 3 authors...
Credit scoring is an important process for peer-to-peer (P2P) lending companies as it determines whether loan applicants are likely to default. The aim of most credit scoring models is to minimize the classification error rate, which implies that all classification errors bear the same cost; however, in reality, there is a significant cost-sensitive problem in credit scoring methods. Therefore, in this paper, a new cost-sensitive logistic regression credit scoring model based on a multi-objectiv...
#1Lili Yuan (CAS: Chinese Academy of Sciences)H-Index: 2
#2Jianping LiXiaolei (CAS: Chinese Academy of Sciences)H-Index: 20
Last. Dengsheng Wu (CAS: Chinese Academy of Sciences)H-Index: 12
view all 5 authors...
Meta-syntheses from experts’ judgements and quantitative metrics are two main forms of evaluation. But they both have limitations. This paper constructs a framework for mapping the evaluation results between quantitative metrics and experts’ judgements such that they may be solved. In this way, the weights of metrics in quantitative evaluation are objectively obtained, and the validity of the results can be testified. Weighted average percentile (WAP) is employed to aggregate different experts’ ...
1 CitationsSource
#1Kao-Yi Shen (CCU: Chinese Culture University)H-Index: 10
#2Hiroshi Sakai (Kyushu Institute of Technology)H-Index: 13
Last. Gwo-Hshiung Tzeng (NTPU: National Taipei University)H-Index: 67
view all 3 authors...
In the recent years, various statistical and computational intelligence or machine learning techniques have contributed to the progress of automation or semiautomation for measuring consumer credit scoring in the banking sector. However, most of the Taiwanese commercial banks still rely on seasoned staffs’ judgments on making the final approvals or rejections. To enhance the understanding and transparency of a decision support system (or model) that can assist bank staffs on making their consume...
3 CitationsSource
#1Yi Qu (CAS: Chinese Academy of Sciences)
#2Pei Quan (CAS: Chinese Academy of Sciences)
Last. Yong ShiH-Index: 42
view all 4 authors...
Abstract Bankruptcy prediction has long been a significant issue in finance and management science, which attracts the attention of researchers and practitioners. With the great development of modern information technology, it has evolved into using machine learning or deep learning algorithms to do the prediction, from the initial analysis of financial statements. In this paper, we will review the machine learning or deep learning models used in bankruptcy prediction, including the classical ma...
2 CitationsSource
#1Tajul Rosli Razak (University of Nottingham)H-Index: 4
#2Jonathan M. Garibaldi (University of Nottingham)H-Index: 34
Last. Daniele Soria (University of Westminster)H-Index: 13
view all 5 authors...
In recent years, researchers have become increasingly more interested in designing an interpretable Fuzzy Logic System (FLS). Many studies have claimed that reducing the complexity of FLSs can lead to improved model interpretability. That is, reducing the number of rules tends to reduce the complexity of FLSs, thus improving their interpretability. However, none of these studies have considered interpretability and complexity from human perspectives. Since interpretability is of a subjective nat...
1 CitationsSource
#1Sait Gül (Beykent University)H-Index: 2
#2Özgür Kabak (ITU: Istanbul Technical University)H-Index: 13
Last. Y. Ilker Topcu (ITU: Istanbul Technical University)H-Index: 13
view all 3 authors...
Abstract Credit rating is a process for building a classification system for credit lenders to characterize current or potential credit borrowers. By such a process, financial institutions classify borrowers for lending decision by evaluating their financial and/or nonfinancial performances. Recently, use of social media data has emerged an important source of information. Accordingly, social media data can be very useful in evaluating companies' credibility when financial or non-financial asses...
2 CitationsSource
#1Sait Gül (Beykent University)H-Index: 2
#2Özgür Kabak (ITU: Istanbul Technical University)H-Index: 13
Last. Y. Ilker Topcu (ITU: Istanbul Technical University)H-Index: 13
view all 3 authors...
Credit lenders utilize credit rating approaches to provide a classification system for characterizing credit borrowers. In order to measure the borrowers’ credibility, that is, ability and willingness to repay the debt, there are many financial and non-financial criteria that should be considered. The basic aim of this study is to propose a multiple-criteria credit rating approach that integrates different kinds of information and represents the borrowers’ credibility as a distribution among all...
3 CitationsSource
#1Yajiao Tang (University of Toyama)H-Index: 2
#2Junkai Ji (University of Toyama)H-Index: 5
Last. Yuki Todo (Kanazawa University)H-Index: 9
view all 6 authors...
Nowadays, credit classification models are widely applied because they can help financial decision-makers to handle credit classification issues. Among them, artificial neural networks (ANNs) have been widely accepted as the convincing methods in the credit industry. In this paper, we propose a pruning neural network (PNN) and apply it to solve credit classification problem by adopting the well-known Australian and Japanese credit datasets. The model is inspired by synaptic nonlinearity of a den...
8 CitationsSource