Treatment of missing data using neural networks and genetic algorithms

Abstract

This paper introduces a method aimed at approximating missing data in a database using a combination of genetic algorithms and neural networks. The proposed method uses genetic algorithm to minimise an error function derived from an auto-associative neural network. An investigation on using the proposed method to accurately approximate missing data as the number of missing cases within a single record increases is conducted. Multi layer perceptron (MLP) and radial basis function (RBF) neural networks are employed. Results obtained using RBF are found to be better than those from the MLP. Results from a combination of both MLP and RBF are found to be better than those obtained using either MLP or RBF individually.