Blind Classification

r

#1

Folks,

I need an algorithm reference which can help me in classifying data on the basis of the patterns.
For example,
I have 100 rows and want to classify this into 10 or more patterns without any supervision.
One more, like identifying the distinct customer behaviour on buying things and associating which have common patterns.

It looks complicated but it’s not.
Also share the R code if possible.

Thanks in advance.


#2

Hi,

Basically it is a clustering task. You can use K Means clustering and specify the desired no. of clusters which is 10 in your case. You can read more about clustering here.


#3

Thanks a ton,

I literally appreciate your suggestion
Piyush as I mentioned classify data on the basis of the pattern but K-mean does not fits for this condition and apology if I had not mentioned the problem properly, actually I don’t have any specific value for cluster like K=10 or something I want machine to do it on own.

Will look forward for more suggestions.
Thanks in advance


#4

scikit-learn python library has implementations of many clustering (at least 6) algorithms - http://scikit-learn.org/stable/modules/clustering.html