Using Neural Networks with Limited Data to Estimate Manufacturing Cost

Neural networks were used to estimate the cost of jet engine components, specifically shaftsand cases. the neural network process was compared with results produced by the currentconventional cost estimation software and linear regression methods. due to the complex natureof the parts and the limited amount of information available, data expansion techniques such asdoubling-data and data-creation were implemented. sensitivity analysis was used to gain anunderstanding of the underlying functions used by the neural network when generating the costestimate. even with limited data, the neural network is able produced a superior cost estimate ina fraction of the time required by the current cost estimation process. when compared to linearregression, the neural networks produces a 30% higher r value for shafts and 90% higher rvalue for cases. compared to the current cost estimation method, the neural network produces acost estimate with a 4.7% higher r value for shafts and a 5% higher r value for cases. thissignificant improvement over linear regression can be attributed to the neural network ability tohandle complex data sets with many inputs and few data points.