A small & experienced Polish software company invented & developed an advanced tool called N-Predictor to predict the future behaviour of systems the exact mathematical models of which can hardly be found, e.g. complex social, economical, natural & technical phenomena. N-Predictor employs feed-forward artificial neural networks with special technology of distributing computations over usual PCs in a local area network. The company is looking for a commercial agreement with technical assistance.
N-Predictor is a dedicated, advanced prediction tool, based on neural networks. It can be used to describe the future behaviour of systems the exact mathematical models of which can hardly be found. Thus, one can think of different complex social, economical, natural and technical phenomena. N-Predictor employs the technology of feed-forward artificial neural networks, trained in several stages by means of improved back-propagation algorithm. A trained network is a system that reacts on particular input signals in an appropriate way, and thus it can be a model of some phenomenon or manufacturing process, predicting its future behaviour. Some new training parameters have been introduced and they are computed in real time, thus allowing to control the entire training process. Contrary to some other software, N-Predictor doesn't offer a wide variety of network architecture, training methods etc. However, this is not a disadvantage, since it is extremely difficult for humans to adjust such free parameters. N-Predictor does it automatically using a special algorithm RDO (Reduced Data Optimisation) developed by the Polish company and proved in practical applications. Since the training of neural networks is usually a time-consuming process and for some complex problems it may take even hundreds of hours, the efficient use of neural software often needs very fast computers, which of course implies high cost. For this reason N-Predictor employs a special technology of distributing computations over usual PCs in a local area network. As shown on the picture below, this is done by a supervising module communicating directly with an Oracle database and with so-called NNA clients (Neural Network Advisors), installed on particular workstations. They are given their jobs to do and they are permanently controlled by the supervising module. The communication is based on D-COM technology. Innovative Aspects: A new method of supervised network training based on the RDO (Reduced Data Optimisation) algorithm developed by the company. Main Advantages: Distributing computations over usual PCs in a local area network allows to avoid the necessity of employing expensive servers and greatly reduces the total hardware cost.