Which of the following statements about the K-means algorithm are correct?

	1.	K-means will always give the same clustering result regardless of the initialization of the centroids.
	2.	A standard way of initilizaing K-means is to set all the centroids, mu1 ro muk, to be a vector of zeros.
	3.	To avoid K-means getting stuck at a bad local optima, we should try using multiple randon initialization.
	4.	The centroids in the K-means algorithm may not be any observed data points.
	5.	The K-means algorithm can converge to different final clustering results, depending on initial choice of representatives.