What is a centroid in the k-means clustering algorithm? A centroid is a randomly chosen data point (vector) expected to be reasonably close to the geometric center of a cluster. A centroid is the geometric center of a cluster. A centroid is the data point (vector) closest to the geometric center of a cluster. A centroid is another name for a cluster. QUESTION 17 Which expression, when applied to the same string, in general, finds more string fragments? string.split("\t ") # One space string.split(" ") # One space string.split() string.split(" ") # Two spaces QUESTION 18 How do you understand the statement "Correlation does not imply causation"? If two variables are highly correlated, changing one of them does not necessarily change the other. If two variables are highly correlated, changing one of them always changes the other. If two variables are unrelated, changing one of them may change the other. If the change of one variable always leads to the change of the other, then they are not correlated.