I have an Unsupervised problem where user's Credit Card payment data is given for each month for various users for one year. One of the feature in the data having "User Id". For most of the User, this User Id having entry for 12 months (one for each month) and their each month data is given like account balance, credit limit, one time payment, cash advance, etc. I need to create cluster of the users based on their spending habit.
I think since it is credit card payment monthly data so we can not just sum the data for each user and then take the average of it and then try to build the model on it so I was thinking to build the K-Means clustering model for each month's data, For ex, For Jan month one model using K-Means, for Feb month one other model using K-Means, like this till Dec month. So their will be total 12 models(each one would be using K-means) in the end for each month. Then whatever the clusters which most number of months are giving we can consider those many no. of clusters for this unsupervised problem and try to make business sense out of it. Is this approach correct ? If not, then how this type of problem can be used for creating clusters where you have data month-wise. Please provide your guidance. Thank you.
