ÖZET Özellikle son yyllarda artan nüfusa paralel olarak kentle?me ve teknolojik geli?melerin giderek artmasynyn önemli sonuçlaryndan biri olan gürültü ayny zamanda önemli bir çevre, ya?am ve sa?lyk sorunudur. Bu sorunun kentsel mekânda en önemli etkeni kent içi trafik ve bundan kaynaklanan gürültüdür. Bu çaly?mada trafikten kaynaklanan gürültünün tahmin edilmesi amacyyla Yapay Sinir A?lary (YSA) yöntemi ile çe?itli modeller geli?tirilmi?tir. YSA modellerinin olu?turulmasynda çe?itli kaynaklardan elde edilen veriler kullanylmy?tyr. Modellemede A?yr ta?yt sayysy (ta?yt/saat), Hafif ta?yt sayysy (ta?yt/saat), toplam saatlik trafik (ta?yt/saat) ve hyz (km/saat) girdi parametresi olarak, trafik kaynakly gürültü (Leq dB(A)) ise çykty parametresi olarak kullanylmy?tyr. Geli?tirilen YSA modeli ile trafi?inin sebep oldu?u gürültünün tahmini sonuçlary; klasik model sonuçlary (Ölçülen) ile kar?yla?tyrylmy? ve sonuçlaryn uyum içerisinde oldu?u görülmü?tür. Matematiksel modellerin yanynda YSA yakla?ymynyn trafikten kaynaklanan gürültünün tahmininde uygun bir yöntem olarak kullanylabilece?i belirlenmi?tir. Anahtar Kelimeler:

ABSTRACT Noise which is one of the important results of gradually increase of urbanization and technologic developments especially in recent years, as parallel with the increasing population also is important problem of environment, life and health. The important factor of the problem in urban areas is inner city traffic and noise caused by that traffic. In this study, aim of the predict of the noise caused by the traffic, various models have been developed with the Artificial Neural Network (ANN). The data obtained from various sources have been used creating the ANN models. In the model as an input data Heavy Vehicle Number (Vehicle/Hour), Light Vehicle Number (Vehicle/Hour), Total Traffic per Hour (vehicle/hour) and Speed (Km/Hour) were used and as an output data Noise caused by the traffic (Leq dB(A)) was used. Developed ANN model and predicted results of noise caused by the traffic were compared with the classical model results (measured) and it has been seen that the results are acceptable. Predicting the noise caused by the traffic it has been determined that instead of the mathematical methods the ANN model can be used. Key words: