Limitation of k-means clustering
The k-means algorithm is designed to be a simple and fast algorithm. Because of the intentional simplicity in its design, it comes with the following limitations:
- The biggest limitation of k-means clustering is that the initial number of clusters has to be predetermined.
- The initial assignment of cluster centers is random. This means that each time the algorithm is run, it may give slightly different clusters.
- Each data point is assigned to only one cluster.
- k-means clustering is sensitive to outliers.
