I think you could choose n samples randomly(and taking care of how well they could symbol the whole dataset). After getting the clusters, ...
確定! 回上一頁