# Can I perform KMeans on a bimodal data?

I am preparing a dataset for KMean clusters. But a series of data appears to be bimodal:

My question is:

Can I perform KMeans on a bimodal data? If not, what kind of transformation can I perform to make the distribution more symmetrical?

Thanks!

• I guess you have to project the modalities onto a common feature space – Upendra01 Dec 31 '18 at 5:12
• Isn't the entire point of any clustering procedure to identify multiple modes?? – whuber Dec 31 '18 at 16:57

There are no restrictions to MacQueen's classic $$K$$-means clustering algorithm that require the data to be unimodal. So long as you can provide an initial partition of the items into K clusters (with bimodal data, perhaps $$K=2$$ initially?), and can calculate the distance between each observation and the centroid for each cluster at each step, you should be able to proceed without any problems.