1 Comment
This paper will apply “STEAMS” methodology on Chocolate Science. The science will mainly address how the antioxidants in chocolate help reduce free radical formation. Free radicals, atoms with an odd number of electrons, damage blood vessels when oxidized by LDL which consequently increases the risk of heart disease (Technology and Engineering). Data was collected on 20+ chocolate ingredient nutrition factors from 60+ different types of chocolate but were missing the Cocoa%. AI Neural Network algorithm was utilized to impute the missing Cocoa%. The hyperbolic tangent activation function was used to create the hidden layer. In order to overcome the Neural over-fit issue, definitive screening design (DSD) DOE technique was used to optimize the AI Neural algorithm. The optimal Neural setting can improve validation fitness R-Square by more than 20%. Based on the optimized neural model, Chocolate Type and Vitamin C are the highest predictors of estimating Cocoa%. Because fruit is high in Vitamin C, there could be further health benefits from dark fruit chocolate. This may indicate the potential to evaluate a 4th Chocolate Type: Fruit Chocolate – which may be healthier than Dark Chocolate. However, commercial fruit chocolate adds a lot of sugar, and Vitamin C is destroyed after processing.
|