A Modular Neural Network Architecture with Additional Generalization Abilities for High Dimensional Input Vectors

iii Abstract In this project a new modular neural network is proposed The basic building blocks of the architecture are small multilayer feedforward networks trained using the Backpropagation algorithm The structure of the modular system is similar to architectures known from logical neural networks The new network is not fully connected and therefore the number of weight connections is much less than in a monolithic multilayer Perceptron The suggested training algorithm works in two stages and is easy to implement in parallel Due to the used modular structure the training is very quick for large input vectors The modular architecture is designed to combine two di erent approaches of generalization known from connectionist and logical neural networks this enhances the generalization ability which is especially signi cant for a high dimensional input space An object oriented implementation of the proposed model was written to sim ulate the behaviour The evaluation using di erent real world data sets showed that the new archi tecture is very useful for high dimensional input vectors For certain domains the learning speed as well as the generalization performance in the modular system is signi cantly better than in a monolithic multilayer feedforward networkIn this project a new modular neural network is proposed The basic building blocks of the architecture are small multilayer feedforward networks trained using the Backpropagation algorithm The structure of the modular system is similar to architectures known from logical neural networks The new network is not fully connected and therefore the number of weight connections is much less than in a monolithic multilayer Perceptron The suggested training algorithm works in two stages and is easy to implement in parallel Due to the used modular structure the training is very quick for large input vectors The modular architecture is designed to combine two di erent approaches of generalization known from connectionist and logical neural networks this enhances the generalization ability which is especially signi cant for a high dimensional input space An object oriented implementation of the proposed model was written to sim ulate the behaviour The evaluation using di erent real world data sets showed that the new archi tecture is very useful for high dimensional input vectors For certain domains the learning speed as well as the generalization performance in the modular system is signi cantly better than in a monolithic multilayer feedforward network