Clustering on CLINC150
0.39ARIAgentic clustering
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| Agentic clusteringk setting=Discover k, predicted number of clusters (ˆk)=169, total cost ($)=$100 + $232026.05 | 0.39 | 0.91 | 78 | |
| LLM-embedding+k-meansk setting=Given k, predicted number of clusters (ˆk)=150, total cost ($)=$0.392026.05 | 0.28 | 0.87 | 70 | |
| Agentic clusteringk setting=Given k, predicted number of clusters (ˆk)=150, total cost ($)=$100 + $222026.05 | 0.28 | 0.88 | 72 | |
| BERTopick setting=Given k, predicted number of clusters (ˆk)=150, total cost ($)=$02026.05 | 0.26 | 0.82 | 65 | |
| BERTopick setting=Discover k, predicted number of clusters (ˆk)=166, total cost ($)=$02026.05 | 0.26 | 0.82 | 70 | |
| SBERT+k-meansk setting=Given k, predicted number of clusters (ˆk)=150, total cost ($)=$02026.05 | 0.21 | 0.83 | 66 | |
| ClusterLLMk setting=Given k, predicted number of clusters (ˆk)=150, total cost ($)=$1.532026.05 | 0.21 | 0.85 | 68 | |
| Huang & Hek setting=Discover k, predicted number of clusters (ˆk)=766, total cost ($)=$100 + $442026.05 | 0.13 | 0.83 | 59 | |
| LDAk setting=Given k, predicted number of clusters (ˆk)=150, total cost ($)=$02026.05 | 0.06 | 0.56 | 30 | |
| TopicGPTk setting=Discover k, predicted number of clusters (ˆk)=14, total cost ($)=$100 + $662026.05 | 0.05 | 0.46 | 15 |