Credit Rating Method with Heterogeneous Information

Corporate credit rating is a very important issue in finance field. A lot of methods such as neural networks, genetic algorithm and support vector machine have been proposed to solve this problem. The credit rating is a complex problem which includes some determinate criteria and other uncertain criteria associating with human judgement which may be vague or linguistic. Therefore, it includes both quantitative value and qualitative value in credit rating. Furthermore, even for the same kind of determinate or uncertain criteria, or in other words, for the same quantitative or qualitative criteria, the assessment domain and scale are also diverse. Some traditional methods transform all the evaluation domain and scale to a uniform one. Accordingly, it may lead to the loss of information so much as the final total departure of the assessment result. A method dealing with heterogeneous information proposed by F. Herrera and L. Martinez et al. is a good solution for this problem which includes various assessment domain and scale. Based on the above, we take the corporate credit rating process as a multi-criteria evaluation problem with heterogeneous information in this paper. And we propose a corporate credit rating method based on multi-criteria evaluation model with heterogeneous information on 2-tuple fuzzy linguistic model. And we give a case study of an auto-manufacture corporate credit rating. The case study shows that the method is feasible for corporate credit rating.