Home
sonyahanson edited this page Aug 24, 2014
·
4 revisions
Welcome to the tossr wiki!
THE SALAD PROFESSIONALS.
Our goal:
- Provide users with automated yet personalized salads.
- Customer satisfaction guaranteed by optimal price solutions.
- Easy user input is important to us! The system should learn from the user's tastes.
How are we going to achieve that:
-
Input format:
- Allergies (hard restriction)
- Desired price range
- Chopped (optional)
- Size
-
Learning from the user's preferences:
- Preferred ingredients
- Preferred combinations
- Will perturb salad with ingredients not used for a long time
- Will provide user with a training set to test both full salads and simple ingredient combinations
-
Training sets will be used to learn generally good combinations:
- Gathered from the internet recipes
- Gathered from store menus
- From insider knowledge of REAL PEOPLE!
-
The structure of the salad has three parts, which are then optimized according to taste, price and the individual:
-
Leaf (1)
-
Toppings (n)
-
Dressing (1)
Possible algorithms to implement:
- Genetic algorithm
- Ant colony optimization algorithm (swarming algorithms in general)
- Simulated annealing
- Recommendation algorithm (alternating least squares, collaborative filtering) (http://www.slideshare.net/srowen/big-practical-recommendations-with-alternating-least-squares)
- Bayesian network