the relationship between neural networks and dea-r (case study: companies stock exchange)

evaluate the performance of companies on the stock exchange using non-parametric methods is very important. dea and dea-r with the strategies for piecewise linear frontier production function and use of available data, assess the stock company. in this study, using a neural network algorithm dea and dea-r is suggested to classify the first companies in the stock exchange; secondly, using the cover models in the nature of input in technology and constant returns to scale non-decreasing scale performance on each floor with propagation neural network is calculated. thirdly, neural network training and repetition, scale efficiency is determined at the end of a functional study is presented on the company's stock.