“…[22] The authors' previous work discusses the application of binary search (BS), Linear Regression, and Neural Networks to improve the computational efficiency of predicting damage initiation. [23] This previous work, however, considered only two possible models, analyzed only a numeric prediction, and used an inherently limited training data set. This paper will show that accuracy can be significantly increased by improving the training data set to better represent the physical reality, analyzing a combination of numeric and discrete predictions, and better formulating the problem statement to reflect the true objective of the damage initiation criteria.…”