Clustering on MASSIVE Intent
0.62ARIAgentic clustering
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| Agentic clusteringk setting=Given k, predicted number of clusters (ˆk)=60, total cost ($)=$100 + $222026.05 | 0.62 | 0.81 | 69 | |
| Agentic clusteringk setting=Discover k, predicted number of clusters (ˆk)=72, total cost ($)=$100 + $232026.05 | 0.6 | 0.83 | 69 | |
| ClusterLLMk setting=Given k, predicted number of clusters (ˆk)=60, total cost ($)=$1.532026.05 | 0.47 | 0.77 | 59 | |
| BERTopick setting=Given k, predicted number of clusters (ˆk)=60, total cost ($)=$02026.05 | 0.45 | 0.73 | 58 | |
| BERTopick setting=Discover k, predicted number of clusters (ˆk)=67, total cost ($)=$02026.05 | 0.43 | 0.73 | 57 | |
| LLM-embedding+k-meansk setting=Given k, predicted number of clusters (ˆk)=60, total cost ($)=$0.392026.05 | 0.4 | 0.73 | 51 | |
| SBERT+k-meansk setting=Given k, predicted number of clusters (ˆk)=60, total cost ($)=$02026.05 | 0.39 | 0.71 | 52 | |
| Huang & Hek setting=Discover k, predicted number of clusters (ˆk)=172, total cost ($)=$100 + $442026.05 | 0.37 | 0.77 | 48 | |
| LDAk setting=Given k, predicted number of clusters (ˆk)=60, total cost ($)=$02026.05 | 0.17 | 0.44 | 29 | |
| TopicGPTk setting=Discover k, predicted number of clusters (ˆk)=15, total cost ($)=$100 + $662026.05 | 0.1 | 0.47 | 31 |