Short Text Clustering on StackOverflow
83.3AccuracyRSTC
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| RSTC2023.05 | 83.3 | 74.11 | |
| DECAMPComment=Proposed2021.04 | 79.7 | — | |
| RSTC-Ccontrastive learning=class-wise, pseudo-labels=SAOT2023.05 | 78.33 | 70.28 | |
| SCCLAveraging=Average over five random runs2021.03 | 75.5 | 74.5 | |
| K-means_IC2023.05 | 74.96 | 70.27 | |
| DECComment=Baseline2021.04 | 74.7 | — | |
| DECE2EComment=Proposed2021.04 | 74.1 | — | |
| LapDMM-TComment=Baseline2021.04 | 72.8 | — | |
| LapDMM-WComment=Baseline (using add. info)2021.04 | 71 | — | |
| SCCL2023.05 | 70.83 | 69.21 | |
| HAC-SDAlgorithm=Hierarchical Agglomerative Clustering2021.03 | 64.8 | 59.5 | |
| SBERT(k-means)2023.05 | 60.55 | 51.79 | |
| DEC-SIFComment=Baseline (using add. info)2021.04 | 59.8 | — | |
| Self-TrainAlgorithm=SIF + Autoencoder2021.03 | 59.8 | 54.8 | |
| Self-Train2023.05 | 59.38 | 52.81 | |
| TF-IDF2023.05 | 58.52 | 59.02 | |
| TF-IDFFeature dimension=1500, Algorithm=K-means2021.03 | 58.4 | 58.7 | |
| RSTC-OTpseudo-labels=traditional OT2023.05 | 56.77 | 60.17 | |
| STC²-LPI2017.01 | 51.14 | — | |
| STC2-LPI2023.05 | 51.14 | 49.1 | |
| STC²-LE2017.01 | 51.13 | — | |
| STC2Comment=Baseline (using add. info)2021.04 | 51.1 | — | |
| STCCAlgorithm=Word2Vec + CNN + K-means2021.03 | 51.1 | 49 | |
| GMMComment=Baseline2021.04 | 50 | — | |
| RecNNPooling=Ave.2017.01 | 40.79 | — | |
| RecNNPooling=Top+Ave.2017.01 | 40.45 | — | |
| STC²-AE2017.01 | 40.05 | — | |
| STC²-LSA2017.01 | 35.81 | — | |
| Para2vec2017.01 | 32.55 | — | |
| RSTC-Icontrastive learning=instance-wise, clustering=k-means2023.05 | 31.31 | 28.66 | |
| K-meansRepresentation=TF-IDF2017.01 | 20.31 | — | |
| BoWFeature dimension=1500, Algorithm=K-means2021.03 | 18.5 | 14 | |
| BOW2023.05 | 17.92 | 13.21 | |
| K-meansRepresentation=TF2017.01 | 13.51 | — | |
| RecNNPooling=Top2017.01 | 13.13 | — | |
| SkipVecType=Bi2017.01 | 9.59 | — | |
| SkipVecType=Combine2017.01 | 9.34 | — | |
| SkipVecType=Uni2017.01 | 8.79 | — | |
| DECComment=Baseline2021.04 | — | 75.3 | |
| DEC-SIFComment=Baseline (using add. info)2021.04 | — | 54.8 | |
| DECAMPComment=Proposed2021.04 | — | 75.6 | |
| DECE2EComment=Proposed2021.04 | — | 75 | |
| K-meansInput=TF2017.01 | — | 7.81 | |
| K-meansInput=TF-IDF2017.01 | — | 15.64 | |
| LapDMM-TComment=Baseline2021.04 | — | 64.1 | |
| LapDMM-WComment=Baseline (using add. info)2021.04 | — | 64.5 | |
| Para2vec2017.01 | — | 27.86 | |
| RecNNPooling=Top2017.01 | — | 9.9 | |
| RecNNPooling=Ave.2017.01 | — | 40.58 | |
| RecNNPooling=Top+Ave.2017.01 | — | 40.21 | |
| SkipVecMode=Uni2017.01 | — | 2.24 | |
| SkipVecMode=Bi2017.01 | — | 2.89 | |
| SkipVecMode=Combine2017.01 | — | 2.72 | |
| STC2Variant=AE2017.01 | — | 38.22 | |
| STC2Variant=LSA2017.01 | — | 34.38 | |
| STC2Variant=LE2017.01 | — | 49.03 | |
| STC2Variant=LPI2017.01 | — | 49.08 | |
| STC2Comment=Baseline (using add. info)2021.04 | — | 49 |