# Financial Natural Language Processing on FinDATA

Source: [Financial Natural Language Processing on FinDATA](<https://www.sota2.com/research/sota/financial-natural-language-processing-on-findata>)

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

Data updated: 2026-02-26T06:06:02.070Z

Evaluation of general Financial Natural Language Processing capabilities on the FinDATA dataset, measured using metrics like TSA, SC, NC, NAD, FSRL, and CD.

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 | Financial Natural Language Processing |
| Dataset (including any stated split) | FinDATA |
| Metric | TSA |
| Unit | SCALAR |
| Direction | Higher is better |
| Results loaded | 12 of 12 |
| 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

- [TSA](<https://www.sota2.com/research/sota/financial-natural-language-processing-on-findata.md?metric=TSA>): Higher is better; unit: SCALAR.

- [SC](<https://www.sota2.com/research/sota/financial-natural-language-processing-on-findata.md?metric=SC>): Higher is better; unit: PERCENT.

- [NC](<https://www.sota2.com/research/sota/financial-natural-language-processing-on-findata.md?metric=NC>): Higher is better; unit: PERCENT.

- [NAD](<https://www.sota2.com/research/sota/financial-natural-language-processing-on-findata.md?metric=NAD>): Higher is better; unit: PERCENT.

- [FSRL](<https://www.sota2.com/research/sota/financial-natural-language-processing-on-findata.md?metric=FSRL>): Higher is better; unit: PERCENT.

- [CD](<https://www.sota2.com/research/sota/financial-natural-language-processing-on-findata.md?metric=CD>): Higher is better; unit: PERCENT.

## Recorded results

| Method | TSA | Result date | Report paper |
| --- | --- | --- | --- |
| FinancialBERT | 0.25 | Not recorded | 68fa6f691e209131cea4569f |
| BERT-cased | 0.232 | Not recorded | 68fa6f691e209131cea4569f |
| Y-FinBERT | 0.2275 | Not recorded | 68fa6f691e209131cea4569f |
| MTL | 0.2159 | Not recorded | 68fa6f691e209131cea4569f |
| MTL | 0.2156 | Not recorded | 68fa6f691e209131cea4569f |
| MTL | 0.2151 | Not recorded | 68fa6f691e209131cea4569f |
| MTL | 0.2083 | Not recorded | 68fa6f691e209131cea4569f |
| MTL | 0.2077 | Not recorded | 68fa6f691e209131cea4569f |
| BERT-uncased | 0.2069 | Not recorded | 68fa6f691e209131cea4569f |
| P-FinBERT | 0.2054 | Not recorded | 68fa6f691e209131cea4569f |

### FinancialBERT

TSA: 0.25. Result ID: 6990ad6e3a6c314849025f4c.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6f691e209131cea4569f.md>)

Recorded evaluation configuration: {"Learning approach":"STL"}. This record does not assert independent reproduction.

### BERT-cased

TSA: 0.232. Result ID: 6990ad6e3a6c314849025f4a.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6f691e209131cea4569f.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa65c31e209131cea0c4ed.md>)

[Implementation](<https://github.com/google-research/bert>)

Recorded evaluation configuration: {"Learning approach":"STL"}. This record does not assert independent reproduction.

### Y-FinBERT

TSA: 0.2275. Result ID: 6990ad6e3a6c314849025f4d.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6f691e209131cea4569f.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa67f01e209131cea193ee.md>)

Recorded evaluation configuration: {"Learning approach":"STL"}. This record does not assert independent reproduction.

### MTL

TSA: 0.2159. Result ID: 6990ad6e3a6c314849025f54.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6f691e209131cea4569f.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6f691e209131cea4569f.md>)

[Implementation](<https://github.com/EdisonNi-hku/MTL4Finance>)

Recorded evaluation configuration: {"Learning approach":"MTL","Backbone":"P-FinBERT","MTL Subset":"Only Sentiment"}. This record does not assert independent reproduction.

### MTL

TSA: 0.2156. Result ID: 6990ad6e3a6c314849025f50.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6f691e209131cea4569f.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6f691e209131cea4569f.md>)

[Implementation](<https://github.com/EdisonNi-hku/MTL4Finance>)

Recorded evaluation configuration: {"Learning approach":"MTL","Backbone":"P-FinBERT","MTL Subset":"w/o FSRL"}. This record does not assert independent reproduction.

### MTL

TSA: 0.2151. Result ID: 6990ad6e3a6c314849025f4f.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6f691e209131cea4569f.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6f691e209131cea4569f.md>)

[Implementation](<https://github.com/EdisonNi-hku/MTL4Finance>)

Recorded evaluation configuration: {"Learning approach":"MTL","Backbone":"P-FinBERT","MTL Subset":"Full FinDATA"}. This record does not assert independent reproduction.

### MTL

TSA: 0.2083. Result ID: 6990ad6e3a6c314849025f53.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6f691e209131cea4569f.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6f691e209131cea4569f.md>)

[Implementation](<https://github.com/EdisonNi-hku/MTL4Finance>)

Recorded evaluation configuration: {"Learning approach":"MTL","Backbone":"P-FinBERT","MTL Subset":"w/o Number"}. This record does not assert independent reproduction.

### MTL

TSA: 0.2077. Result ID: 6990ad6e3a6c314849025f51.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6f691e209131cea4569f.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6f691e209131cea4569f.md>)

[Implementation](<https://github.com/EdisonNi-hku/MTL4Finance>)

Recorded evaluation configuration: {"Learning approach":"MTL","Backbone":"P-FinBERT","MTL Subset":"w/o CD"}. This record does not assert independent reproduction.

### BERT-uncased

TSA: 0.2069. Result ID: 6990ad6e3a6c314849025f4b.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6f691e209131cea4569f.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa65c31e209131cea0c4ed.md>)

[Implementation](<https://github.com/google-research/bert>)

Recorded evaluation configuration: {"Learning approach":"STL"}. This record does not assert independent reproduction.

### P-FinBERT

TSA: 0.2054. Result ID: 6990ad6e3a6c314849025f4e.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6f691e209131cea4569f.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa66441e209131cea0f5a8.md>)

[Implementation](<https://github.com/google-research/bert>)

Recorded evaluation configuration: {"Learning approach":"STL"}. This record does not assert independent reproduction.

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