Introducing the Partial Clustering Problem
Byron J. Gao · 2023
In this paper we introduce the partial clustering problem. Given o objects, select $p \le o$ of them such that the selected objects form an optimal clustering with respect to a given cost function f. The problem can find various applications where due to capacity constraints, only part of the data objects can be selected, and the selection criteria are based on how good a clustering the partial data can form. Partial clustering generalizes clustering, and is significantly harder involving nested combinatorial optimization. For the introduced problem, we also propose generic algorithms without specifications of the cost function, and perform preliminary experiments for initial verification.