Clustering users based on apps he uses


#1

I have prepared a dataset in which there are users and android apps installed on his device.
I have another dataset which has android apps and their score. A score is a number decided manually, which depicts how good an app is.
I am trying to cluster users into good,bad and average, based on the apps installed on his phone.

How should i approach this problem?

P.S: Sorry for the vague explanation.


#2

Hi @dvik,

The first thing I would suggest you to do is to define the threshold of “goodness” for a person is. Decide in how would you map score of a mobile app to score of a person.

Once you have decided this, the problem can be converted into a supervised learning task, with labels as your person “goodness” score.

Then you have to create features which would help you map people to their score (for example, gender of a person, or his/her usage statistics of the app), which can help you classify these people into categories of the goodness.

Hope this helps.


#3

It also depends on what metrics you are using for clustering them.