A0179
Title: Mis-specification analysis of a pH acceleration model
Authors: S-T Tseng - National Tsing-Hua University (Taiwan) [presenting]
Abstract: Shelf-life prediction of highly reliable nano-sol products is an interesting research topic. Recently, a pH acceleration model has been proposed in the literature for the time evolution of nano-particle distributions. There are two approaches for modelling the particle-size distribution, by using either a parametric approach (mixture-normal distribution) or a non-parametric (distribution free) approach. Their shelf-life predictions, however, have significant differences. Therefore, quantifying the effects of model mis-specification of particle size distribution on the shelf-life prediction of nano-sol products turned out to be a challenging issue. The accuracy and precision of the shelf-life prediction were obtained analytically if the true particle-size follows a mixture-normal distribution, but wrongly treated as a distribution free model. For the given set of data, the shelf-life prediction may be under-estimated up to 13.66\%, while its relative variation may be inflated up to 13.35 times by using a non-parametric approach. In addition, simulations show that when the sample size and the measurement times are small, these effects are moderately large and cannot be neglected.
