A continuously learning neural network

In real world we are always faced with incomplete and "fuzzy" data. We humans quite often make everyday decisions based on our limited knowledge. In this space we present a continuously learning neural network that will predict the future performance of a certain stock based on it's past performance. We will use this tool to help us determine whether to buy or sell that stock.

The stock we choose to study is that of Goldman Sachs (stock symbol GS). The network we designed is a five layer network. It consists of one input, one output, and three hidden layers for a total of five layers. The algorithm that makes this neural network learn is the well known back propagation algorithm that employs the gradient heuristic to organize in ways that improves its performance over time. Figure below describes the network assembly.

Network