METHOD AND SYSTEM FOR PREDICTING CONSTITUENT YIELDS IN TOBACCO SMOKE USING A MULTIVARIATE REGRESSION MODEL

摘要

The concentrations or yields of a first set of components in a particular tobacco smoke, such as the Hoffmann analytes, are predicted on the basis of a statistical model. This model is derived from a multivariate regression analysis that relates the concentrations of the first set of components across a range of tobacco smokes to the yields of a second set of components. Typically the second set of components includes gases and other substances such as carbon monoxide, whose concentration can be determined relatively easily. Thus the (unknown) concentrations of Hoffmann analytes in a particular tobacco smoke can be predicted by first measuring the yields of the second set of components in the particular tobacco smoke to be investigated, and then using the multivariate regression model to predict the concentrations of the first set of components from the measured concentrations of the second set of components.