[Tutor] K-means and Python Beginner - The clusters shown in the graph seems wrong

[DL Neil]

On 20/04/20 4:16 AM, tsz ying tsang via Tutor wrote:
> Dear Tutor,
> I have recently signed up an online class on coursera - Data Science - K-mean Clustering, and i had submitted an assignment about using k-means to build a model distinguishing the genuine and fake banknotes by the "banknote authentication dataset" on OpenML (our assignment used a simplified one as attached)
> I have written some code on Python and plt graph (k=2). But the clusters I get is quite different from most of my online classmates. May I know what i did wrong in my code that can't generate the same cluster.
> How to re-run the k-means to make sure the cluster is stable?
> Thanks so much for help first
> The below is my classmate's one.


Regret to advise that this Discussion List does not accept attachments 
unless they are simple-text.

That sounds like an interesting course. Why don't you post this question 
on the Discussion List for the applicable week of your course - that's 
one of the reasons it is there! You are likely to find a class-colleague 
or a Teaching Assistant who is familiar with the actual problem-set, who 
will help you...

That said, there are a few people 'here' who may be able to help, once 
we can 'see' the problem...

May I further suggest, that in addition to any code and data you may 
care to send, that you first explain your approach (in English rather 
than Python). Yes, the difference may be explained by a coding error, 
but it is usually the clustering criteria which account for such 
differences!
-- 
Regards =dn