Bu çalışmada bulanık mantık ve Yapay Sinir Ağlarını (YSA) bir arada kullanan bir bulanık-YSA metodu oluşturulmuştur. Bu metodun amacı çok boyutlu veri uzayında 'belirgin' bir şekilde lineer olarak birbirinden ayrılmış kümeler elde etmektir. 'Belirgin' kümeleri elde etmek için, bulanık c-ortalamalar metodu ile hesaplanan küme merkezleri koordinatları, YSA metotlarının girdisini oluşturmuştur. Geleneksel küme analizi ve YSA metotlarının aksine giriş verisi bulanık küme merkezleri koordinatlarına göre türetilmiştir. Bulanık-YSA metodunda en iyi sonucu elde etmek için, İleri Beslemeli Geri Yayılımlı Ağ (İBGYSA), Genelleştirilmiş Regresyon Ağı (GRYSA) ve Radyal Tabanlı Fonksiyon Ağları (RTFA) olmak üzere üç farklı YSA metodu kullanılmıştır. Bulanık-YSA metodu, literatürde küme analizi metotlarının performansının test edilmesi için sıkça kullanılan iris verisi üzerinde denenmiş ve % 94'lük bir performans göstererek küme analizi problemlerinde kullanılabileceği gösterilmiştir. Bulanık-YSA modeli, bulanık c-ortalamalar metodu ve YSA tabanlı olduğundan, sonuçları bulanık c-ortalamalar metodu ve YSA 'nın küme analizinde kullanılan türü Kohonen YSA (KYSA) ile karşılaştırdığımızda, bulanık-YSA modeli hem bulanık c- ortalamalar metodu (% 90) hem de KYSA 'dan (% 88.67) daha iyi sonuç vermiştir. Bulanık-YSA metodu, Türkiye iklim bölgelerinin belirlenmesi için korelasyonlardan oluşan matrisin kullanıldığı küme analizlerinde, Ward metodu ve KYSA metoduna göre daha başarılı bulunmuştur. Yağış rejimi bölgelerinin belirlendiği küme analizlerinde ise Ward metodu bulanık-YSA metoduna göre daha iyi sonuç vermiştir.

A neural network can approximate a function, but it is impossible to interpret the result in terms of natural language. The fusion K of neural networks and fuzzy logic in neuro-fuzzy models provide learning as well as readability. Control engineers find this useful, because the models can be interpreted and supplemented by process operators. A neural network can model a dynamic plant by means of a nonlinear regression in the discrete time domain. The result is a network, with adjusted weights, which approximates the plant. It is a problem, though, that the knowledge is stored in an opaque fashion; the learning results in a (large) set of parameter values, almost impossible to interpret in words. Conversely, a fuzzy rule base consists of readable if-then statements that are almost natural language, but it cannot learn the rules itself. The two are combined in neuro-fuzzy systems in order to achieve readability and learning ability at the same time. The obtained rules may reveal insight into the data that generated the model, and for control purposes, they can be integrated with rules formulated by control experts (operators). This article introduces a method which is a mixture of the fuzzy c-means clustering method and the artificial neural networks (ANN) used in one system. The goal of this method is to obtain clearly separated clusters in high dimensional data space. In order to obtain clear 'obvious' clusters, the cluster centers calculated by the fuzzy c-means method is used as input for the artificial neural networks. Unlike conventional cluster analysis methods and artificial neural networks, the input data is generated according to the coordinates of cluster centers calculated by the fuzzy c-means method. Three different artificial neural network techniques, namely, the Feed Forward Back Propagation (FFBP), the Generalized Regression Neural Networks (GRNN) and the Radial Basis Function-Based Neural Networks (RBFNN) are used in the neuro-fuzzy method in order to achieve better performance. The performance of the proposed method was illustrated using the Iris data set which is commonly used in the international literature. We tested the performance all three cluster analysis methods which were the Ward's method, the Koho-nen-ANN and a neuro-fuzzy method developed in this study for defining the precipitation regime regions and the climate zones. The Irrs data which is commonly used in the literature for determining the performance of the clustering methods was used to test the performance of the neuro-fuzzy method. The performance of the neuro-fuzzy method was 94% and this result showed that the neuro-fuzzy method can be used in the cluster analyses. This result is better than Ward's method result (89.33%) which is commonly used for defining climate zones, the Ko-honen-ANN (88.67%) and the fuzzy Kohonen-ANN (91.33%) which is a mixture of fuzzy set theory and ANN. Thus, an improvement had been made the results of the fuzzy c-means clustering method and the Kohonen-ANN with the created neoro-fuzzy method that uses fuzzy the c-means method and ANN methods. According to the stability analyses of the clustering methods applied in this study were showed that these 3 methods can be applied in the cluster analyses which have different number of stations. The results of the stability analysis of these 3 methods were so close to each other but Kohonen-ANN had showed a bit better performance than the Ward's method and the Kohonen-ANN. A matrix created from the correlation coefficients of the meteorological data were used as input for cluster analysis methods. In the international literature, there is no similar approach like thecreation of this correlation matrix. Due to appearance of "sub-climate zones" when the correlation coefficients were used as input, sub-precipitation regime regions" and "sub-climate zones" of Turkey were manage to demonstrate by maps. The generated neuro-fuzzy method's results were more consistent on determination of climate zones considering the influence of large scale pressure systems and upper air circulation, location of transition regions, topography, exposure, continentally, the controls of physical geography and orograpy. The ward's method results didn 't used due to unexpected transitition regions. According to the results of the generated neuro-fuzzy method, there is seven main climate region and 15 sub-climate regions in Turkey. The 'sub-climate' regions of the main climate zones are revealed for the first time in the international literature.