Here we propose a differentiable cluster quality measure, and an associated clustering algorithm to optimize it.
Schmitzer-Torbert, N., Jackson, J., Henze, D., Harris, K. & Redish, A. D. Quantitative measures of cluster quality for use in extracellular recordings.
Finally, automatic methods determine the number of modules completely automatically, usually by iterating over several parameter settings and selecting the one that optimizes some criterion of cluster quality.
Schmitzer-Torbert, N., Jackson, J., Henze, D., Harris, K. & Redish, A.D. Quantitative measures of cluster quality for use in extracellular recordings.
Additionally, three evolving rules are designed to simultaneously boost the cluster quality and the reconstruction error during the iterative updates.
Normally, image segmentation algorithms like fuzzy clustering use spatial information along with the color information to improve the cluster quality.
Truncating centroids to the most significant terms (e.g., ) hardly decreases cluster quality while achieving a significant speedup of the reassignment step (see references in Section 16.6 ).
Measures for cluster quality can range from the stability of clusters among several resamplings of the dataset27, the balance between cluster tightness and separateness (as measured by cluster validity indices28), or the optimal functional enrichment of the modules22.
Comparison with related methods is made in terms of number of data points used, computational time and a cluster quality measure.
The assessment of the results is often carried out with the help of cluster quality measures which take into account the compactness and separation of the clusters.
Our experiments show that the NP-hard model yields a better cluster quality at the cost of lower coverage, due to the fact that our instances cannot be completely solved by our algorithms.
