Study of the Relevance of Polynomial Order in Takagi-Sugeno Fuzzy Inference Systems Applied in Diagnosis Problems

摘要: Fuzzy Logic has been implemented successfully for different kind of problems. However, there is an opportunity these methods to be improved in the realm classifications The present paper focused a specific application classification problems, diagnosis systems, this problem consists training intelligent system learn relationship between symptoms and diagnosis, problems are usually based powerful non-linear example Modular Neural-Networks or complex hybrids models, however, applied Type-1 Takagi Sugeno Systems (TSK) but analyzing improvement their performance by increasing order polynomial. conventional Takagi-Sugeno aggregation first-order polynomial it interesting observe effect increase polynomial, TSK Diagnosis evaluated accuracy obtained ten benchmark dataset UCI Dataset Repository, diseases difficult levels.