algorithm - Python - Stratified Sampling with MiniBatch k-means -
i'm trying cluster million objects, each have varying length of datapoints, less 100. features date of observations , id value of each object(let's runner's(name) , times in different races). want run minibatch k-means on data, want algorithm take stratified samples based on third feature, let's u.s. state runner from. there way implement such sampling within minibatchkmeans function?
if not, there way take stratified samples , pass them function somehow? thought this, seems if took stratified samples , passed them base k-means algorithm, wouldn't able aggregate samples , 1 label each object. advice?
Comments
Post a Comment