Text Classification
Text Classification is the task of assigning a sentence or document an appropriate category. The categories depend on the chosen dataset and can range from topics.
Text Classification problems include emotion classification, news classification, citation intent classification, among others. Benchmark datasets for evaluating text classification capabilities include GLUE, AGNews, among others.
In recent years, deep learning techniques like XLNet and RoBERTa have attained some of the biggest performance jumps for text classification problems.
( Image credit: Text Classification Algorithms: A Survey )
Papers
Showing 76–100 of 3635 papers
All datasetsMTEBAG NewsDBpediaR8TREC-620NEWSUK Key Stage ReadabilityOhsumedYahoo! AnswersMRClimabenchNewsDiscourse
Benchmark Results
| # | Model | Metric | Claimed | Verified | Status |
|---|---|---|---|---|---|
| 1 | ST5-XXL | Accuracy | 73.42 | — | Unverified |
| 2 | ST5-XL | Accuracy | 72.84 | — | Unverified |
| 3 | ST5-Large | Accuracy | 72.31 | — | Unverified |
| 4 | Ada Similarity | Accuracy | 70.44 | — | Unverified |
| 5 | SGPT-5.8B-nli | Accuracy | 70.14 | — | Unverified |
| 6 | ST5-Base | Accuracy | 69.81 | — | Unverified |
| 7 | SGPT-5.8B-msmarco | Accuracy | 68.13 | — | Unverified |
| 8 | MPNet-multilingual | Accuracy | 67.91 | — | Unverified |
| 9 | GTR-XXL | Accuracy | 67.41 | — | Unverified |
| 10 | SimCSE-BERT-sup | Accuracy | 67.32 | — | Unverified |
| # | Model | Metric | Claimed | Verified | Status |
|---|---|---|---|---|---|
| 1 | Mistral-Small-24B + CAPO | Error | 15.7 | — | Unverified |
| 2 | ToWE-SG | Error | 14 | — | Unverified |
| 3 | Qwen2.5-32B + CAPO | Error | 12.93 | — | Unverified |
| 4 | Llama-3.3-70B + CAPO | Error | 11.2 | — | Unverified |
| 5 | Seq2CNN with GWS(50) | Error | 9.64 | — | Unverified |
| 6 | Char-level CNN | Error | 9.51 | — | Unverified |
| 7 | SVDCNN | Error | 9.45 | — | Unverified |
| 8 | VDCN | Error | 8.67 | — | Unverified |
| 9 | Balanced+bi-leaf-RNN | Error | 7.9 | — | Unverified |
| 10 | CCCapsNet | Error | 7.61 | — | Unverified |
| # | Model | Metric | Claimed | Verified | Status |
|---|---|---|---|---|---|
| 1 | Seq2CNN(50) | Error | 2.77 | — | Unverified |
| 2 | Char-level CNN | Error | 1.55 | — | Unverified |
| 3 | SWEM-concat | Error | 1.43 | — | Unverified |
| 4 | FastText | Error | 1.4 | — | Unverified |
| 5 | VDCN | Error | 1.29 | — | Unverified |
| 6 | CCCapsNet | Error | 1.28 | — | Unverified |
| 7 | Balanced+bi-leaf-RNN | Error | 1.2 | — | Unverified |
| 8 | BERT large UDA | Error | 1.09 | — | Unverified |
| 9 | M-ACNN | Error | 1.07 | — | Unverified |
| 10 | EXAM | Error | 1 | — | Unverified |
| # | Model | Metric | Claimed | Verified | Status |
|---|---|---|---|---|---|
| 1 | DeBERTa | Accuracy | 98.45 | — | Unverified |
| 2 | C-BERT (ESGNN + BERT) | Accuracy | 98.28 | — | Unverified |
| 3 | ESGNN | Accuracy | 98.23 | — | Unverified |
| 4 | RoBERTaGCN | Accuracy | 98.2 | — | Unverified |
| 5 | BERT | Accuracy | 98.17 | — | Unverified |
| 6 | SGNN | Accuracy | 98.09 | — | Unverified |
| 7 | ERNIE 2.0 | Accuracy | 98.04 | — | Unverified |
| 8 | DistilBERT | Accuracy | 97.98 | — | Unverified |
| 9 | Our Model* | Accuracy | 97.8 | — | Unverified |
| 10 | ALBERTv2 | Accuracy | 97.62 | — | Unverified |
| # | Model | Metric | Claimed | Verified | Status |
|---|---|---|---|---|---|
| 1 | TM-Glove | Error | 9.96 | — | Unverified |
| 2 | byte mLSTM7 | Error | 9.6 | — | Unverified |
| 3 | SWEM-aver | Error | 7.8 | — | Unverified |
| 4 | DELTA (CNN) | Error | 7.8 | — | Unverified |
| 5 | Capsule-B | Error | 7.2 | — | Unverified |
| 6 | STM+TSED+PT+2L | Error | 7.04 | — | Unverified |
| 7 | GRU-RNN-GLOVE | Error | 7 | — | Unverified |
| 8 | MPAD-path | Error | 6.2 | — | Unverified |
| 9 | VLAWE | Error | 5.8 | — | Unverified |
| 10 | C-LSTM | Error | 5.4 | — | Unverified |
| # | Model | Metric | Claimed | Verified | Status |
|---|---|---|---|---|---|
| 1 | LinearSVM+TFIDF | Accuracy | 93 | — | Unverified |
| 2 | RoBERTaGCN | Accuracy | 89.5 | — | Unverified |
| 3 | SSGC | Accuracy | 88.6 | — | Unverified |
| 4 | SGC | Accuracy | 88.5 | — | Unverified |
| 5 | SGCN | Accuracy | 88.5 | — | Unverified |
| 6 | RMDL (15 RDLs) | Accuracy | 87.91 | — | Unverified |
| 7 | Sparse Tensor Classifier | Accuracy | 87.3 | — | Unverified |
| 8 | GraphStar | Accuracy | 86.9 | — | Unverified |
| 9 | NABoE-full | Accuracy | 86.8 | — | Unverified |
| 10 | Text GCN | Accuracy | 86.34 | — | Unverified |
| # | Model | Metric | Claimed | Verified | Status |
|---|---|---|---|---|---|
| 1 | ELECTRA + ANN | F1 | 99.6 | — | Unverified |
| 2 | ERNIE + ANN | F1 | 99.4 | — | Unverified |
| 3 | XLNet + ANN | F1 | 99.2 | — | Unverified |
| 4 | RoBERTa + ANN | F1 | 98.7 | — | Unverified |
| 5 | Longformer + ANN | F1 | 93.9 | — | Unverified |
| 6 | BERT + ANN | F1 | 90.5 | — | Unverified |
| 7 | ALBERT + ANN | F1 | 79.7 | — | Unverified |
| 8 | BERT | F1 | 75 | — | Unverified |
| 9 | DistilBERT | F1 | 74.4 | — | Unverified |
| 10 | XLNet | F1 | 74 | — | Unverified |
| # | Model | Metric | Claimed | Verified | Status |
|---|---|---|---|---|---|
| 1 | RoBERTaGCN | Accuracy | 72.8 | — | Unverified |
| 2 | Our Model* | Accuracy | 69.4 | — | Unverified |
| 3 | SSGC | Accuracy | 68.5 | — | Unverified |
| 4 | SGC | Accuracy | 68.5 | — | Unverified |
| 5 | SGCN | Accuracy | 68.5 | — | Unverified |
| 6 | Text GCN | Accuracy | 68.36 | — | Unverified |
| 7 | GraphStar | Accuracy | 64.2 | — | Unverified |
| 8 | ApproxRepSet | Accuracy | 64.06 | — | Unverified |
| 9 | REL-RWMD k-NN | Accuracy | 58.74 | — | Unverified |
| 10 | CNN+Lowercased | Accuracy | 36.2 | — | Unverified |
| # | Model | Metric | Claimed | Verified | Status |
|---|---|---|---|---|---|
| 1 | BERT-ITPT-FiT | Accuracy | 77.62 | — | Unverified |
| 2 | DRNN | Accuracy | 76.26 | — | Unverified |
| 3 | DELTA (HAN) | Accuracy | 75.1 | — | Unverified |
| 4 | EXAM | Accuracy | 74.8 | — | Unverified |
| 5 | DNC+CUW | Accuracy | 74.3 | — | Unverified |
| 6 | ULMFiT (Small data) | Accuracy | 74.3 | — | Unverified |
| 7 | CCCapsNet | Accuracy | 73.85 | — | Unverified |
| 8 | SWEM-concat | Accuracy | 73.53 | — | Unverified |
| 9 | FastText | Accuracy | 72.3 | — | Unverified |
| 10 | Seq2CNN(50) | Accuracy | 55.39 | — | Unverified |
| # | Model | Metric | Claimed | Verified | Status |
|---|---|---|---|---|---|
| 1 | DeBERTa | Accuracy | 90.21 | — | Unverified |
| 2 | RoBERTaGCN | Accuracy | 89.7 | — | Unverified |
| 3 | ERNIE 2.0 (optimized) | Accuracy | 89.53 | — | Unverified |
| 4 | RoBERTa | Accuracy | 89.42 | — | Unverified |
| 5 | ERNIE 2.0 | Accuracy | 88.97 | — | Unverified |
| 6 | BERT | Accuracy | 86.94 | — | Unverified |
| 7 | ALBERTv2 | Accuracy | 86.02 | — | Unverified |
| 8 | DistilBERT | Accuracy | 85.31 | — | Unverified |
| 9 | SSGC | Accuracy | 76.7 | — | Unverified |
| # | Model | Metric | Claimed | Verified | Status |
|---|---|---|---|---|---|
| 1 | CliReBERT (P0L3/clirebert_clirevocab_uncased) | Evaluation Macro F1 | 0.65 | — | Unverified |
| 2 | ClimateBERT (climatebert/distilroberta-base-climate-f) | Evaluation Macro F1 | 0.64 | — | Unverified |
| 3 | BERT (google-bert/bert-base-uncased) | Evaluation Macro F1 | 0.61 | — | Unverified |
| 4 | CliSciBERT (P0L3/cliscibert_scivocab_uncased) | Evaluation Macro F1 | 0.61 | — | Unverified |
| 5 | SciBERT (allenai/scibert_scivocab_cased) | Evaluation Macro F1 | 0.59 | — | Unverified |
| 6 | DistilRoBERTa (distilbert/distilroberta-base) | Evaluation Macro F1 | 0.58 | — | Unverified |
| 7 | SciClimateBERT (P0L3/sciclimatebert) | Evaluation Macro F1 | 0.58 | — | Unverified |
| 8 | RoBERTa (FacebookAI/roberta-base) | Evaluation Macro F1 | 0.57 | — | Unverified |
| # | Model | Metric | Claimed | Verified | Status |
|---|---|---|---|---|---|
| 1 | Human (Post-Rec.) (Spangher et al., 2021) | macro F1 | 73.69 | — | Unverified |
| 2 | MT-Mac (Spangher et al., 2021) | macro F1 | 63.46 | — | Unverified |
| 3 | MT-Mic (Spangher et al., 2021) | macro F1 | 61.89 | — | Unverified |
| 4 | RL-IP/TT (Choubey et al., 2021) | macro F1 | 57 | — | Unverified |
| 5 | Document LSTM + Document encoding (Choubey et al., 2020) | macro F1 | 54.4 | — | Unverified |
| 6 | CRF Fine-grained (Choubey et al., 2020) | macro F1 | 52.9 | — | Unverified |
| 7 | Human (Blind) (Spangher et al., 2021) | macro F1 | 46.18 | — | Unverified |
| 8 | Feature-based (SVM) (Choubey et al., 2020) | macro F1 | 38.3 | — | Unverified |
| # | Model | Metric | Claimed | Verified | Status |
|---|---|---|---|---|---|
| 1 | 1-6 BertGCN | Accuracy | 96.6 | — | Unverified |
| 2 | GraphStar | Accuracy | 95 | — | Unverified |
| 3 | Our Model* | Accuracy | 94.6 | — | Unverified |
| 4 | SSGC | Accuracy | 94.5 | — | Unverified |
| 5 | SGC | Accuracy | 94 | — | Unverified |
| 6 | SGCN | Accuracy | 94 | — | Unverified |
| 7 | Text GCN | Accuracy | 93.56 | — | Unverified |
| 8 | TM-Glove | Accuracy | 89.14 | — | Unverified |