Assigning weights to a variable to Score a product




Can anyone suggest me an statistical approach of assigning weights to multiple product attributes to achieve a score against products which is a indicator of product rank in terms of revenue and conversion(Orders/ visit).




Can you provide more details, what do you want to achieve? Which kind of products are we talking about? The question is not clear as of now.



@kunal Sorry for that.
I am working on an E commerce project so product can be anything from a toothpaste to a house. We want to show the best products to a viewer in each category(in order to increase Revenue and order conversion) , for that we are scoring each of the products based on some parameters and using that score we sort our products. Currently the score is calculated based on assigning weights on these parameters(Say 10% to var1, 20 % to var2 and so on), and these weights are determined based on our gut feeling. I want to optimize these weights to revamp our scoring mechanism using some data oriented approach to achieve a more relevant score.

Hope now the problem is clear.




This looks like a classic recommendor system problem. Do you have user ratings / feedback for products / categories, which can be used to solve this?

You can search for user-user collaborative recommendor system and item-iterm collaborative recommendor systems to start with and then decide the best way depending on the data which is available with you.

You might also find this course on Coursera helpful:



@kunal : Thanks for the reply, but currently we are not going into personalization at a user level. What I actually want to do is optimized weights on some parameters which will help me calculate a score for each product, through which I will sort my products in a certain category irrespective of a user perspective. I am looking for approaches to find a solution for this kind of optimization problem.


Hi Kunal,

I am also working on similar area where online users will rate, give feedback, ask questions (and many other activities) for different products / brands / categories and we are interested to calculate users’ trustworthiness/loyalty/credit point based on several attributes. Is there any method to give proper weightage to different attributes and then find trustworthiness/loyalty/credit point of users ?

You have talked about user-user collaborative recommendor system and item-iterm collaborative recommendor systems. Will these methods apply here ? If so can you please share some good materials. If not, what your suggestions on methods to be used for this one ?



Did you get solution to your weight problem? I am also looking for the same solution.



I am also going through same situation but I want to give weights in HR industry. So can you please give me advice how to give weightage of particular variables?

Like in My problem I got score but now I want to give weightage depending upon Variable like need to give 60% weightage to Title and location score those are primary variable and else are secondary variable so Can you please tell me How to give weightage on variable?

If you have any material or link please provide me for reference.

Thank you in advance


Hi Prem,

For employee score you can use Glassdoor api.
You can look for features like comparing salary, skillset, designation and I guess your company must be having a criteria to rate employee yearly, so you can use those variables also.
then you can use classification or clustering to assign a score.

Hope it helps.


@Rizwann Thanks for kind reply I resolved my issue by below link


Hi, the link is no longer available,
can you write few words on the topic (like, which algorithm you got into using?)