Abstract
Artificial
Neural Networks (ANN) are important data processing algorithms which are used
for solving nonlinear problems. Through classical approaches, mathematical
infrastructure and complex equations in ANN are difficult to understand.
Interactive and multimedia-based courseware has the potential to overcome these
difficulties. In this study, a web based educational courseware for ANN was
developed to provide an effective and efficient learning environment so that
the difficulties can be overcome. This interactive courseware was also enriched
with animations and text-based course contents. In addition to this, the
effects of ANN parameters’ changes were observed directly through graphical
results. In this way, users can easily understand the fundamentals and working
mechanism of ANN. Without using any commercial libraries, the courseware was
developed with ASP.NET, an object-oriented programming language. The courseware
supports file formats such as XML, TXT, and CSV so that it can co-operate with
other software. “Balance and Scale” data set was used to evaluate the
performance of the courseware. 0.9918 accuracy, 1 specificity and 1 sensitivity
values were achieved. When this study is compared to previous studies,
improvements in terms of visuality, understandability and interactivity can
clearly be identified.