Journal of applied mathematics & informatics
- Volume 26 Issue 5_6
- /
- Pages.861-876
- /
- 2008
- /
- 2734-1194(pISSN)
- /
- 2234-8417(eISSN)
FUZZY ERROR MATRIX IN CLSSIFICATION PROBLEMS
- Kannan, S.R. (Department of Mathematics, Gandhigram Rural University) ;
- Ramathilagam, S.R. (Sree Sastha Institute of Engg. And Technology)
- Published : 2008.09.30
Abstract
This paper concerns a new method called Fuzzy Supervised Method for error matrix, the method has developed based on Adoptive Neuro- Fuzzy Inference Systems(ANFIS). For the performance point of view initially the new method tested with trial data and then this paper applies the proposed method with real world problems. So that this paper generated 1000 random error matrices in programming language [R] and then it tests the new proposed method for the error matrices. The results of Fuzzy Supervised Method given in terms of Kappa Index and Congalton Accuracy Indexes, and performance of Fuzzy Supervised Method has evaluated by using Pearson's test.