CODEWITHSUNDEEP
algorithmmathlogicaverageweighted-average
anvd
anvd

Reputation: 4057

Algorithm to calculate a general score and each criteria independently

Imagine something like booking reviews, where we have for each review some scores by criteria and a general rating.

For the sake of simplicity i will use letters, but imagine A like comfort, B localization, and so on. Available rating is (1-4)

                  A   B   C   D 
hotel_review1  |  1   2   3   4 
hotel_review2  |  2   2   3   3

                   A   B   C   D
 user_weighing  |  2   3   2   3

For the global rating I have this (just a basic weighing avg):

enter image description here

For the avg of each criteria, I am doing an arithmetic avg and after that, multiply the result by the user_weighing:

A - ((1+2)/2)*2/4 = 0.75
B - ((2+2)/2)*3/4 = 1.5
C - ((3+3)/2)*2/4 = 1.5
D - ((4+3)/2)*3/4 = 2.625
                    -----
                    6.375 / 4 = 1.60

So, my problem is that i need a general score, in my case 2.55. However if i look for each score independently, there is a big difference between 2.55 and 1.60. There is a better way of doing what I am looking for?

Upvotes: 1

Views: 111

Answers (1)

samgak
samgak

Reputation: 24427

As mentioned by Juhana in the comments, you need to divide by the sum of the weights, which is 10 (2 + 3 + 2 + 3), and not by the number of weights, which is 4:

A: ((1+2)/2)*2/10 =  0.3
B: ((2+2)/2)*3/10 =  0.6
C: ((3+3)/2)*2/10 =  0.6
D: ((4+3)/2)*3/10 = 1.05
                   -----
                    2.55

However it is a more simple calculation requiring less operations if you leave the divisions by the number of reviews (2) and the sum of weights (10) until the end:

A: (1+2)*2 =  6.0
B: (2+2)*3 = 12.0
C: (3+3)*2 = 12.0
D: (4+3)*3 = 21.0
            -----
             51.0 / (2*10) = 2.55

Upvotes: 1

Related Questions