affinity estimation, forest-based metric, unsupervised clustering forest, pseudo-leaf-splitting (PLS)
This paper presents an unsupervised cluste-ring random-forest-based metric for affinity estimation in large and high-dimensional data. The criterion usedfor node splitting during forest construction can handle rank-deficiency when measuring cluster compactness. The binary forest-based metric is extended to continuous metrics by exploiting both the common traversal path and the smallest shared parent node.
Yi, Yunai; Sun, Diya; Li, Peixin; Kim, Tae-Kyun; Xu, Tianmin; and Pei, Yuru
"Unsupervised random forest for affinity estimation,"
Computational Visual Media: Vol. 8:
2, Article 6.
Available at: https://dc.tsinghuajournals.com/computational-visual-media/vol8/iss2/6