| Autor: |
Femina Bahari T., Sudheep Elayidom M. |
| Předmět: |
|
| Zdroj: |
International Journal of Advanced Research in Computer Science; Jan/Feb2018, Vol. 9 Issue 1, p632-635, 4p |
| Abstrakt: |
In this paper we propose an efficient fuzzy computational framework for customer segmentation model in credit analysis. Normally segmentation methods cannot perform complex analysis so as to obtain the customer segments with high value. If the knowledge of experts in the data domain can be imparted to the generation of segments, this can bring in better results in the performance of classification models. In our approach customer attributes are selected after knowledge experts analysis and are segmented based on the limits set by them on the real numerical values. For each segment we generate the segmentation rules with definition of fuzzy basic linguistic term set. Each linguistic term set is assigned to a fuzzy membership function to generate the segmentation function. Combining the segmentation rules and generated functions the real valued numerical attributes are converted to fuzzified values in the interval [0, 1]. Both linguistic and numeric information are aggregated by a series of computations and a 2-tuple linguistic value is generated for each attribute in the database. The same term after a series of computations can be used in many decision making problems as it suffers no loss of information. [ABSTRACT FROM AUTHOR] |
| Databáze: |
Complementary Index |
| Externí odkaz: |
|
