Loading...

Credit Scoring of Commercial Loan Applicants in Iranian Banking Industry, A Comparative Analysis of Bayesian Approach, Logit, and Neural Networks

Ghanbari, Hamed | 2010

466 Viewed
  1. Type of Document: M.Sc. Thesis
  2. Language: Farsi
  3. Document No: 40416 (44)
  4. University: Sharif University of Technology
  5. Department: Management and Economics
  6. Advisor(s): Zamani, Shiva; Bahramgiri, Mohsen
  7. Abstract:
  8. The development of effective models for classification problems, such as the problem of selecting which credit applicants to accept, has been the subject of intense research for decades. Many static and dynamic methods, ranging from statistical classifiers to decision trees, nearest-neighbor methods, and neural networks, have already been proposed to tackle this problem and to assist decision making in the area of consumer and commercial credit. Given the profusion of modeling and data management techniques, it is often the case that which model has the more appropriate outputs in classification of the same problem. Among the stated methods although the latter, Neural Networks, is powerful pattern recognition techniques and its use for practical problem solving is rather limited due to their intrinsic opaque, black box nature. Recently, Bayesian network classifiers have been introduced in the literature as probabilistic white-box classifiers with the capacity of giving a clear insight into the structural relationships in the domain under investigation. Bayesian Networks (BNs) are an increasingly popular formalism for reasoning and decision-making in problems that involve uncertainty and probabilistic reasoning. BN is a probabilistic graphical model whose nodes represent random variables and connecting arcs represent direct relationships between the variables. Basically, in a BN, by giving some evidences, which stands for observation of some nodes, one want to know the posterior probability of other nodes. A Bayesian net encodes the probability distribution of a set of attributes by specifying a set of conditional independence assumptions together with a set of relationships among these attributes and their related joint probabilities. When used in this way,Bayesian networks result in a powerful knowledgerep resentation formalism based on probability and provide a natural way of dealing with uncertainty and complexity, two recurring topics that have impact across a wide range of knowledge domains. In the process of choosing the useful Bayesian network classifiers for this thesis, I used the results of comparative analysis, which evaluated and contrasted various types of Bayesian network classifiers including Naïve Bayesian, Unrestricted Bayesian, Tree Augmented Naïve Bayesian, and Learning Bayesian Networks. I choose a Tree Augmented Naive Bayesian framework to estimate the posterior probability of being in a certain class given multiple classifiers. Then, loan applicants were classified in accordance with this model. The results of this classification were compared with “Logit” and “Neural networks” approach. I showed that Augmented Naïve Bayesian Network is more appropriate than previously used method for classification of loan applicants in this Bank. I also showed that the results of classification with neural networks were more suitable than the result of BNs.
  9. Keywords:
  10. Credit Scoring ; Credit Rating ; Classification ; Classification ; Bayesian Network ; Neural Network

 Digital Object List

 Bookmark

No TOC