# Part-of-Speech Tagging on Treebank Union (test)

Source: [Part-of-Speech Tagging on Treebank Union (test)](<https://www.sota2.com/research/sota/part-of-speech-tagging-on-treebank-union-test>)

Collection index: [llms.txt](<https://www.sota2.com/research/llms.txt>)

Data updated: 2026-02-26T05:53:38.061Z

Evaluation of Part-of-Speech (POS) Tagging performance on the test split of the Treebank Union dataset, reporting domain-specific metrics for News, Web, and QTB.

A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage.

| Scope | Recorded value |
| --- | --- |
| Task | Part-of-Speech Tagging |
| Dataset (including any stated split) | Treebank Union (test) |
| Metric | News Domain Accuracy |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 6 of 6 |
| Coverage | All results currently returned for this benchmark |

Evaluation settings may differ between results. The original papers and recorded configurations are needed to establish comparability. A shared metric name alone does not establish an identical protocol.

## Metrics

- [News Domain Accuracy](<https://www.sota2.com/research/sota/part-of-speech-tagging-on-treebank-union-test.md?metric=News%20Domain%20Accuracy>): Higher is better; unit: PERCENT.

- [Web Domain Accuracy](<https://www.sota2.com/research/sota/part-of-speech-tagging-on-treebank-union-test.md?metric=Web%20Domain%20Accuracy>): Higher is better; unit: PERCENT.

- [QTB Domain Accuracy](<https://www.sota2.com/research/sota/part-of-speech-tagging-on-treebank-union-test.md?metric=QTB%20Domain%20Accuracy>): Higher is better; unit: PERCENT.

## Recorded results

| Method | News Domain Accuracy | Result date | Report paper |
| --- | --- | --- | --- |
| Our Global (B=8) | 97.77 | Not recorded | 68fa62871e209131ce9f9edd |
| Our Local (B=8) | 97.69 | Not recorded | 68fa62871e209131ce9f9edd |
| Our Local (B=1) | 97.66 | Not recorded | 68fa62871e209131ce9f9edd |
| Linear CRF | 97.6 | Not recorded | 68fa62871e209131ce9f9edd |
| Parsey McParseface | 97.52 | Not recorded | 68fa62871e209131ce9f9edd |
| Ling et al. (2015) | 97.44 | Not recorded | 68fa62871e209131ce9f9edd |

### Our Global (B=8)

News Domain Accuracy: 97.77. Result ID: 698bdb0cecace47ef5592b59.

[Reporting paper](<https://www.sota2.com/research/paper/68fa62871e209131ce9f9edd.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa62871e209131ce9f9edd.md>)

[Implementation](<http://github.com/tensorflow/models/tree/master/syntaxnet>)

Recorded evaluation configuration: {"Normalization Strategy":"Global","Beam Size (B)":8}. This record does not assert independent reproduction.

### Our Local (B=8)

News Domain Accuracy: 97.69. Result ID: 698bdb0cecace47ef5592b58.

[Reporting paper](<https://www.sota2.com/research/paper/68fa62871e209131ce9f9edd.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa62871e209131ce9f9edd.md>)

[Implementation](<http://github.com/tensorflow/models/tree/master/syntaxnet>)

Recorded evaluation configuration: {"Normalization Strategy":"Local","Beam Size (B)":8}. This record does not assert independent reproduction.

### Our Local (B=1)

News Domain Accuracy: 97.66. Result ID: 698bdb0cecace47ef5592b57.

[Reporting paper](<https://www.sota2.com/research/paper/68fa62871e209131ce9f9edd.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa62871e209131ce9f9edd.md>)

[Implementation](<http://github.com/tensorflow/models/tree/master/syntaxnet>)

Recorded evaluation configuration: {"Normalization Strategy":"Local","Beam Size (B)":1}. This record does not assert independent reproduction.

### Linear CRF

News Domain Accuracy: 97.6. Result ID: 698bdb0cecace47ef5592b55.

[Reporting paper](<https://www.sota2.com/research/paper/68fa62871e209131ce9f9edd.md>)

Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction.

### Parsey McParseface

News Domain Accuracy: 97.52. Result ID: 698bdb0cecace47ef5592b5a.

[Reporting paper](<https://www.sota2.com/research/paper/68fa62871e209131ce9f9edd.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa62871e209131ce9f9edd.md>)

[Implementation](<http://github.com/tensorflow/models/tree/master/syntaxnet>)

Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction.

### Ling et al. (2015)

News Domain Accuracy: 97.44. Result ID: 698bdb0cecace47ef5592b56.

[Reporting paper](<https://www.sota2.com/research/paper/68fa62871e209131ce9f9edd.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa627c1e209131ce9f9812.md>)

[Implementation](<https://github.com/wlin12/JNN>)

Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction.

Page 1 of 1. This is an index page, not the full collection.