keyword
cluster indicator matrix
A cluster indicator matrix is a mathematical matrix used in data clustering and machine learning to represent the assignment of data items to distinct groups or clusters. In this representation, rows typically correspond to individual data points or features and columns correspond to specific clusters, with the numerical entries denoting the degree or state of membership for each item. In hard clustering formulations, the matrix is composed of binary or orthogonal indicator vectors where each item has an exclusive non-zero entry assigning it to a single cluster. In soft, relaxed, or fuzzy clustering frameworks, such as nonnegative matrix factorization and spectral clustering, the matrix entries consist of non-negative real values representing probabilities, weights, or continuous latent factors that describe partial cluster membership.
1 item
