{"paper":{"title":"FinBERT: Financial Sentiment Analysis with Pre-trained Language Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"FinBERT adapts a BERT model with financial text to improve sentiment classification on specialized datasets.","cross_cats":["cs.LG"],"primary_cat":"cs.CL","authors_text":"Dogu Araci","submitted_at":"2019-08-27T07:40:48Z","abstract_excerpt":"Financial sentiment analysis is a challenging task due to the specialized language and lack of labeled data in that domain. General-purpose models are not effective enough because of the specialized language used in a financial context. We hypothesize that pre-trained language models can help with this problem because they require fewer labeled examples and they can be further trained on domain-specific corpora. We introduce FinBERT, a language model based on BERT, to tackle NLP tasks in the financial domain. Our results show improvement in every measured metric on current state-of-the-art res"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"Our results show improvement in every measured metric on current state-of-the-art results for two financial sentiment analysis datasets. We find that even with a smaller training set and fine-tuning only a part of the model, FinBERT outperforms state-of-the-art machine learning methods.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"That further pre-training on domain-specific financial corpora plus partial fine-tuning will reliably produce better sentiment classification than general-purpose models or traditional machine-learning baselines when labeled financial data is limited.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"FinBERT adapts BERT to the financial domain and outperforms prior state-of-the-art methods on financial sentiment analysis tasks.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"FinBERT adapts a BERT model with financial text to improve sentiment classification on specialized datasets.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"499bedf3667912b6b9843da21f6663b2295ad14bb3a6c5c91de3ca38d596a765"},"source":{"id":"1908.10063","kind":"arxiv","version":1},"verdict":{"id":"55305194-ebfd-4011-8938-dd39477e3c7d","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-15T20:22:30.685641Z","strongest_claim":"Our results show improvement in every measured metric on current state-of-the-art results for two financial sentiment analysis datasets. We find that even with a smaller training set and fine-tuning only a part of the model, FinBERT outperforms state-of-the-art machine learning methods.","one_line_summary":"FinBERT adapts BERT to the financial domain and outperforms prior state-of-the-art methods on financial sentiment analysis tasks.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"That further pre-training on domain-specific financial corpora plus partial fine-tuning will reliably produce better sentiment classification than general-purpose models or traditional machine-learning baselines when labeled financial data is limited.","pith_extraction_headline":"FinBERT adapts a BERT model with financial text to improve sentiment classification on specialized datasets."},"references":{"count":35,"sample":[{"doi":"10.1007/978-3-319-25343-5_3","year":2016,"title":"Basant Agarwal and Namita Mittal. 2016. Machine Learning Approach for Sentiment Analysis. Springer International Publishing, Cham, 21–45. https: //doi.org/10.1007/978-3-319-25343-5_3","work_id":"211e4061-c85e-45ea-9c00-5487ea1743c5","ref_index":1,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"10.1016/j.eswa.2017.02.002","year":2017,"title":"Fernando Sánchez-Rada, and Carlos A","work_id":"2805d73d-a21e-431c-85a9-66ae1d3690b0","ref_index":2,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2018,"title":"Shallue and Jaehoon Lee and Joe Antognini and Jascha Sohl-Dickstein and Roy Frostig and George E","work_id":"3f2397b9-748a-4ee8-b371-99c0a4f13cdb","ref_index":3,"cited_arxiv_id":"1811.03600","is_internal_anchor":true},{"doi":"10.1016/j.jfds.2017.02.001","year":2016,"title":"Li Guo, Feng Shi, and Jun Tu. 2016. Textual analysis and machine leaning: Crack unstructured data in finance and accounting. The Journal of Finance and Data Science 2, 3 (sep 2016), 153–170. https://d","work_id":"9de23409-dd72-439a-8f28-769462db129f","ref_index":4,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2018,"title":"Universal language model fine-tuning for text classification","work_id":"9990d84d-84b4-4fd4-acb5-da451524d2f4","ref_index":5,"cited_arxiv_id":"1801.06146","is_internal_anchor":true}],"resolved_work":35,"snapshot_sha256":"819ac636916b7226b2487cd63c250fc7f6c121e5e48040a7c6fec7fdc92c8f89","internal_anchors":13},"formal_canon":{"evidence_count":2,"snapshot_sha256":"dd4184825ff3e1bd012511be15e1a0de1d052c497b705b89d516dcee8701bc69"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"}