Neural networks for construction cost estimation: On the positive side neural networks demonstrate many features that are coherent with the nature of cost estimation. Major strengths include their ability to self-organize knowledge from training data, their ability to generate results from incomplete information and their ability to cope with complex relations. these largely resemble the way human estimators develop expertize through experience

The positive side of working with neural networks has to be balanced with an understanding of the difficulties, states the author, and it is hoped that a clear exposition of the strengths and limitations of neural networks will provide researchers with some help on what are unsolved issues in this area.