Knowledge discovery for large data sets using artificial neural network

This paper deals with the application of artificial neural network for extracting knowledge of the predicted income for a prospective employee in US based on the already available database created from the population survey conducted in US. The survey reports the details of demographic and employment related characteristics of around 200,000 instances. A prediction module is developed using backpropagation multilayer perceptron (MLP) neural network to predict the income. A comparative study is undertaken for prediction of income for different learning rates. The validation results show that neural network achieved an acceptable level of performance in terms of maximum error being 0.000009045.