Publications / 2000 Proceedings of the 17th ISARC, Taipei, Taiwan

Fuzzy Neural Network Model for Mark-Up Estimation

Li-Chung Chao, Min Liu
Pages 1-6 (2000 Proceedings of the 17th ISARC, Taipei, Taiwan, ISBN 9789570266986, ISSN 2413-5844)
Abstract:

Mark-up estimation involves many uncertain and complex factors making it difficult to model them using conventional mathematical methods. Artificial neural networks (ANNs), which can adapt themselves to training data, have recently been applied to mark-up estimation. However, there are two major drawbacks of this application; firstly, an ANN system is unable to explain why and how a particular recommendation is made, and secondly the prediction performance of ANNs may not be satisfying in some circumstances or near the boundaries of the training sets. This paper proposes a fuzzy neural network (FNN) model to rectify these drawbacks of ANN for mark-up estimation. The FNN model embodies the fuzzy logic in a neural network as a way of retaining the strengths of both methods. While experts? judgments in linguistic rules are captured by the network structure, neural network parameters are fine-tuned through training with objective quantitative data. Because the results are produced within the scope of clearly stated rules and interpretable parameters, the inference process is now transparent. At the same time, the accuracy also improves. An example is presented to illustrate the performance and applicability of the proposed FNN model in comparison with ANN models.

Keywords: mark-up, neural network, fuzzy logic, fuzzy neural network, construction management.