While preparing for POSSCON, and a possible chance to meet Walter Bender (one of the lead developers of the Sugar software), we are also laying the foundation in completing a new Sugar Activity. Our team member Alex has presented us with a GUI she built using Glade, a program that utilizes PyGTK and XML to produce things like combo boxes and allows for data entry into categories. We also suggested the possibility of including class diagrams in the documentation, but we have not come across the need for many so the number of diagrams will be fairly small.
We also started to think more about the algorithm involved in calculating the score of the user's diet. I mentioned using a regression formula the week before, because the calculation involves a weighted sum. With more research, I found the answer to be just that,
straight-line linear regression. In our case, the variables are as follows:
- weight = the weight of each food group (according to a food pyramid TBD) = (Servings of Group) / (Total Servings)
- numServings = the user will pick a number of servings of a certain food group they have eaten (Alex currently has the measurement of Fists)
- foodGroupScore = sum(weightn * numServingsn) --> where n corresponds to a specific food group as selected by the user
Still, there is a level deeper that may be ignored in this calculation in including the impact each food in each food category has in determining the suggested number of servings of that food group because it is already given by the food pyramid of choice.
No comments:
Post a Comment