Neural networks for business decision making

In the current big-data era, business decision making usually involves mining large datasets for finding hidden patterns which can be used for predictions. Such data analytical tasks are far beyond the capabilities of human experts. Artificial Neural Networks (ANNs) are non-linear models that resemble biological neural networks in structure and learn through training. ANNs learn from examples in a way similar to how the human brain learns. Then ANNs take complex and noisy data as input and make educated guesses based on what they have learned from historical data. This paper presents a new learning algorithm for Higher Order Neural Networks (HONNs) which are ANNs in which the net input to a computational neuron is a weighted sum of its inputs plus products of its inputs. The novel learning algorithm is based on Extreme Learning Machine (ELM) algorithm which randomly chooses hidden layer neurons and analytically determines output weights. The experimental results demonstrate that HONN models with the new algorithm offer significant advantages over standard HONN models and traditional ANNs (including Multilayer Perceptrons and RBF Networks), such as faster training and improved generalization abilities.