[] Clustering algorithm principle of DBSCAN

Density-based clustering algorithm: DBSCAN

  1. It requires two parameters are predefined: "radius of influence" and "Threshold Effect"
  2. Advantages are: kmeans not like the previously determined as the value of k, accommodate nonconventional data, can be used to find outliers (outliers)
  3. The disadvantages are: two parameters still need to be determined in advance
  4. The key to understanding: the development of similar MLM downline or tribal expansion
  5. Schematic:Here Insert Picture Description
  6. DBSCAN visual display
Published 40 original articles · won praise 0 · Views 1704

Guess you like

Origin blog.csdn.net/weixin_44680262/article/details/104702236