True or false: One can use binary values in the k-Means clustering model.

• A. True
• B. False
• C. Not allowed
• D. Not applicable

Answer: (A)

Binary values can be used in k-Means because the method operates on numeric data and relies on distances and averages. In the algorithm, points are assigned to the nearest cluster using a distance measure (typically Euclidean), and then each cluster center is updated to be the mean of its member points across all features. When a feature is binary, its mean within a cluster is simply the proportion of points in that cluster that have a value of 1, so the centroid component lands between 0 and 1. This produces a valid distance calculation to binary points, allowing the algorithm to run and produce clusters. Of course, the centroids become fractional rather than binary, which can be less interpretable for strictly binary data, and for such data there are other methods (like k-modes or alternative distance measures) that may be more appropriate. In practice, using binary data with k-Means is common and workable, especially when you interpret centroid values as frequencies or probabilities.