Layered-CasPer: Layered Cascade Artificial Neural Networks

Abstract

Previous research has demonstrated that constructive algorithms are powerful methods for training feedforward neural networks. The CasPer algorithm is a constructive neural network algorithm that generates networks from a simple architecture and then expands it. The A-CasPer algorithm is a modified version of the CasPer algorithm which uses a candidate pool instead of a single neuron being trained. This research adds an extension to the A-CasPer algorithm in terms of the network architecture - the Layered-CasPer algorithm. The hidden neurons form as layers in the new version of the network structure which results in less computational cost being required. Beyond the network structure, other aspects of Layered-CasPer are the same as A-CasPer. The Layered-CasPer algorithm extension is benchmarked on a number of classification problems and compared to other constructive algorithms, which are CasCor, CasPer, A-CasPer, and AT-CasPer. It is shown that Layered-CasPer has a better performance on the datasets which have a large number of inputs for classification tasks. The Layered-CasPer algorithm has an advantage over other cascade style constructive algorithms in being more similar in topology to the familiar layered structure of traditional feedforward neural networks.