I have a problem where I need to predict a list of objects based on previous history of the usage of the objects. It is a recommendation system in short.
I figured I can use clustering on existing data, and then try to find pattern among the clusters.
For this I came acros scikit-learn library in python, and i think it will work.
But I need to know how I will use one of their clustering algorithms(say MeanShift) , since the examples they provide mostly work on their own datasets provided in the library itself.
So,
How do I organize my data so that I can use the MeanShift class from sklearn.cluster package?
My data points are multidimensional, so will I be able to use sklearn package in the first place? they haven't mentioned any constraints.
If I can cluster multidimensional data points, will I have to do dimensionality reduction? ( I don't know how to do this either, but I am aware of the concept)
I have done some data mining in one of my courses, but these are new waters for me, any help in terms of pointing to resources/tutorials will be appreciated hightly.
Thank you.