Using SAS and .NET

Neural networks are a powerful method for solving complex, "real world", modeling problems when traditional algorithms cannot be formulated. Most neural networks are implemented in programming languages that can represent data structures well such as C/C++, Java or Visual Basic. SAS is very good at processing longititudinal data. This paper will offer techniques and methods through implementation of neural nets using only BASE SAS and SAS macro's to represent intrinsic iterative learning/training steps of neural network algorithms. This SAS program can be used as building blocks to build one's own neural network models or be used as a starting point to study neural networks for SAS programmers. A numeric example for predicting stock prices is presented in this paper using a .NET application framework. Introduction There are different kinds of neural network models. For example, Adaline Madaline, Backpropagation, Bidirectional Associative Memory, Temporal Associative Memory, Brain-State-in-a-box, Counterpropagation, Neocognitron, Adaptive Resonance Theory(ART) etc.. Most research on implementations about neural network models will show that many implementations are done in C++, Java or Visual Basic. These languages have very good abilities to express data structures such as trees, linked lists, queues and stacks. In this paper will focus on a three-layer Feedforward Backpropagation neural net . By being able to design and implement a simple neural net model in Base SAS we can know how to build more complicated neural network models. In addition we will demonstrate this application using a .NET “wrapper” interface that does not use standard SAS external application components such as SAS/Intrnet or Integration Technologies.