{"paper_id":"2311.07598","title":"Multi-Label Topic Model for Financial Textual Data","authors":["Moritz Scherrmann"],"paper_date":"2023-11-10","paper_type":"empirical","paper_url":"https://arxiv.org/pdf/2311.07598v1","page_url":"https://thequant.space/flowcharts/multi-label-topic-model-for-financial-textual-data/","scores":{"empirical_rigor":6.5,"math_complexity":4.5,"quadrant":"Street Traders"},"proposes":{"status":"available","statement":"The paper proposes a BERT‑based multi‑label topic model for financial texts and creates a manually labeled German ad‑hoc announcement dataset with 20 topics.","quote":"I train the model on a new financial multi‑label database consisting of 3,044 German ad‑hoc announcements that are labeled manually using 20 predefined, economically motivated topics.","summary":"This paper presents a multi-label topic model for financial texts like ad-hoc announcements, 8-K filings, finance related news or annual reports. I train the model on a new financial multi-label database consisting of 3,044 German ad-hoc announcements that are labeled manually using 20 predefined…","strategy":null},"assumptions":{"status":"available","items":[{"assumption":"The gold labels are treated as the ground truth for the labeling process.","type":"data","stated":"explicit","quote":"In the prior sections, I assume that the gold labels are the ground truth of the labeling process."},{"assumption":"At least 20 announcements are required for each topic pair to be included in the regression analysis.","type":"statistical","stated":"explicit","quote":"I require at least 20 announcements in the sample for each topic pair, which reduces the number of topic combinations considered in the regression to 116."}]},"datasets":{"status":"available","items":[{"name":"ad‑hoc multi‑label database","provider":null,"kind":"text-news","access":"unknown","period":"1996‑2020","universe":"German ad‑hoc announcements","quote":"After duplicate removal, the final sample of the ad‑hoc multi‑label database consists of 31,771 sentences from 3,044 announcements."}],"known":[{"id":"refinitiv","name":"Refinitiv / LSEG (Datastream, Eikon, Tick History)","provider":"LSEG","kind":"market-prices","access":"commercial","quote":"I gather the market data from Thomson Reuters Datastream.","mentions":2,"used_in":1}],"sample":{"start":null,"end":null,"market":null,"asset_class":null,"frequency":null},"data_resolution":"daily"},"transaction_costs":{"status":"not_applicable","modeled":"n/a","headline_net_of_costs":null,"detail":null,"quote":null,"source":"front_matter","rule":"R3","candidates":{"front_matter":"n/a","evidence_card":"n/a","evidence_card_quoted":false},"meaning":"the paper reports no trading or execution result"},"results":{"status":"available","evaluation_type":"forecast-accuracy","findings":[{"finding":"The best model achieves a macro F1 score of more than 85%, indicating high predictive accuracy.","kind":"predictive-accuracy","quote":"The best model achieves a macro F1 score of more than 85%."},{"finding":"Topics like Large Scale Project, Buyback or Squeeze Out have a positive and statistically significant impact on abnormal returns (2.7, 1.21 and 1.21 percentage points, respectively).","kind":"performance","quote":"Topics like Large Scale Project, Buyback or Squeeze Out have a positive and statistically significant impact on"},{"finding":"Topics such as Law, Profit Warning, Delay, and Bankruptcy Filing have a significant negative influence on abnormal returns (-1.45, -4.04, -4.29 and -8.75 percentage points, respectively).","kind":"performance","quote":"topics like Law, Profit Warning, Delay, or Bankruptcy Filing have a significant negative influence on abnormal returns (-1.45, -4.04, -4.29 and -8.75 percentage points, respectively)."},{"finding":"Stock market reactions are highly topic dependent and also depend on the co‑occurrence of other topics within announcements.","kind":"other","quote":"on single topics also depend on the co‑occurrence of other topics that appear in announcements."}],"reported_performance":null,"out_of_sample":{"method":"unclear","headline_out_of_sample":null}},"independent_evidence":{"status":"none_found","supports":[],"contradicts":[],"qualifies":[],"compared_with":38,"citations":{"cited_by_in_corpus":0,"by_independent_authors":0,"how_cited":{"used_as_baseline":0,"uses_method_or_data":0,"extends":0,"discusses_findings":0,"mentions":0,"unclassified":0},"independent_papers":[],"scope":"Citations between papers in this corpus only; citations from outside it are not counted."}},"implementation":{"status":"available","public":false,"code_url":null,"note":"No repository presented by the authors as their own was found in the paper. Absence here is not proof that no code exists."},"replication":{"status":"none_found","card":null,"note":"No replication of this paper has been run by thequant.space yet."},"extraction":{"text_source":"fulltext","model":"openai/gpt-oss-120b","version":1,"extracted_at":"2026-11-14T18:19:45Z"},"provenance":{"claims":"Everything here is what the authors state in the paper. Nothing was re-run or reproduced by the platform.","quotes":"Assumptions, datasets, cost detail and findings were extracted by a language model. Every quote is verbatim paper text. Where an item has `support`, \"full\" means every figure in it appears in the quoted sentences and \"partial\" means a figure is stated elsewhere in the paper; items with a figure found nowhere in the paper were discarded.","datasets_known":"datasets.known lists well-known vendors and public datasets by a stable id, found by a dictionary over the full text where the sentence says the data is used; it is independent of the model and covers only names in the dictionary.","flags":"transaction_costs.modeled, results.evaluation_type and results.out_of_sample.method are platform classifications (structured-decision models); see https://thequant.space/score-guide/#penalty-flags","comparisons":"supports / contradicts / qualifies are a language model's reading of two papers' quoted findings, restricted to closely related papers that share no author. They are pointers to read, not verified replications, and coverage is partial.","implementation":"code_url is a repository the paper itself presents as its own (rule-scored, ambiguous cases judged by a model), checked for liveness when it was recorded."},"docs":"https://thequant.space/agents/"}