Table 1 Analysis of advantages and disadvantages of clustering algorithms
From: Analysis of influencing factors on excellent teachers' professional growth based on DB-Kmeans method
Algorithm name | Advantages | Disadvantages |
|---|---|---|
DBSCAN | (1) No need specify the number of clusters in advance (2) Outliers and clusters of any shape can be found | (1) When the sample data is large, the clustering convergence time is long (2) When the sample data is quite different, the clustering effect is poor (3) The parameters are complex |
Kmeans | (1) The clustering principle is simple and there are few parameters, so the clustering time is fast (2) There are few parameters, so the process is simple (3) The clustering effect is good and the interpretability is strong | (1) Specify the K value in advance, and the selection is not easy to grasp (2) Generally, only applicable to convex datasets (3) The initial value is completely random, so the results may belong to the local optimal solution |