Akademik Veri Yönetim Sistemi
International Research in Engineering Sciences, DOÇ. DR. MUSTAFA ALTIN, Editör, EĞİTİM YAYINEVİ, Konya, ss.5-26, 2022
Flow estimation is necessary for the operation, protection,
control, management and optimization of water resources
(Noari and Kalin, 2016: 141). Making river flow forecasts not
only help regulate reservoir outflows during times of low flow
of rivers to manage water resources, but also provide warning
of upcoming stages for flood control. These accurate estimates
also help provide accurate information for city planning,
designing water resources projects, designing hydropower
projects, preparing management plans and implementing
practices that reduce the environmental impact of climate
change.For these reasons, accurate estimation of stream or river
flow is very important (Anusree and Varghese, 2016; Besaw et
al., 2010). Many techniques and models have been used in order
to predict the currents correctly and to increase the accuracy of
the predictions made, and the development of these models have
gained importance day by day. Recently, artificial intelligence
methods have been used for hydrological applications such as precipitation-flow modeling, flood forecasting, precipitation
forecasting, water quality modeling (Tayyab et al., 2016:108).
ANNs, whose historical development started in the 1970s,
and whose nearly 30 different models were developed in the
following 10 years are the best work of technology that resembles
the processing system of our brain and is inspired by our nerve
cells. Within the scope of this study, monthly flow values were
estimated by using adaptive neural-based fuzzy inference system
(ANFIS), Long Short-Term Memory (LSTM) and Feed Forward
Neural Network (FFNN) algorithms