
#1
Dec1912, 10:47 AM

P: 6

Hey all,
So the idea is that I'm trying to create a composite ranking system of items that are already in different categories. For example, suppose there are 4 houses that a buyer is choosing from. These 4 houses are ranked 14 in each of the categories such as affordability, location, and design. These categories are then rated Very Important to Not Important (14, respectively). I want to create a final ranking of the housings based on this data. What's the best way to do this? What I've thought of: Pick a house, and multiply its ranking in a category by the importance of that category. Continue to do this for that house for each category, continuing to add to the score. Then divide this score by the maximum score which is the summation of the importance values. Then finally multiply this by 100 to get a percentage match score. Here's the idea: Ʃ(house ranking in a category*importance of that category)/Ʃ(importances) * 100 I feel like this is too simplistic and results in error. What would be a better way? The important thing is coming up with a relevant score value for each house. I've looked at Borda, Condorcet, and Range systems as examples, but I'm not sure what is best. 



#2
Dec1912, 11:54 AM

P: 61

Ʃ(house ranking in a category*importance of that category) does all the work.
The rest is just 'normalization'  scaling for convenience. With this scheme the lower the score the better: A house ranked #1 in all 4 categories would have a score of 10. A house ranked #4 in all 4 categories would have a score of 40. 



#3
Dec1912, 02:05 PM

P: 748




Register to reply 
Related Discussions  
Multiple regression, why use categories  Set Theory, Logic, Probability, Statistics  1  
can collision of two protons create multiple antiprotons ?  High Energy, Nuclear, Particle Physics  2  
Method for ranking multiple items based on a population of rankings?  Set Theory, Logic, Probability, Statistics  2  
9v Batteries Powering Multiple Items  Electrical Engineering  3 