Scholarlink

We present a metric for evaluating the robustness of clusters as information is removed. Utilizing the edit distance between two clusterings as a measure of stability, we test this metric on an Erdos-Renyi random graph and on four different networks representing the connections between patients who experience chronic lower back pain. On all graphs, we utilize both k-means and the Louvain algorithm for cluster identification to test the metric.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

A Measure of the Robustness of Clusters in a Network with No Ground Truth—A Chronic Lower Back Pain Case Study

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Iris Ho,
  • Paul Anderson,
  • Jean Davidson,
  • Jeffrey Lotz,
  • Theresa Migler

摘要

We present a metric for evaluating the robustness of clusters as information is removed. Utilizing the edit distance between two clusterings as a measure of stability, we test this metric on an Erdos-Renyi random graph and on four different networks representing the connections between patients who experience chronic lower back pain. On all graphs, we utilize both k-means and the Louvain algorithm for cluster identification to test the metric.