{"paper_id":"2404.18470","title":"ECC Analyzer: Extract Trading Signal from Earnings Conference Calls using Large Language Model for Stock Performance Prediction","authors":["Yupeng Cao","Zhi Chen","Qingyun Pei","Nathan Jinseok Lee","K. P. Subbalakshmi","Papa Momar Ndiaye"],"paper_date":"2024-04-29","paper_type":"empirical","paper_url":"https://arxiv.org/pdf/2404.18470v2","page_url":"https://thequant.space/flowcharts/ecc-analyzer-extract-trading-signal-from-earnings-conferenc/","scores":{"empirical_rigor":7,"math_complexity":3.5,"quadrant":"Street Traders"},"proposes":{"status":"available","statement":"The paper proposes ECC Analyzer, a novel framework that uses large language models to extract multi‑level information from earnings conference calls for improved volatility prediction.","quote":"... this paper introduces the ECC Analyzer — a novel framework that leverages LLMs for in-depth analysis of ECC data to enhance volatility prediction performance.","summary":"In the realm of financial analytics, leveraging unstructured data, such as earnings conference calls (ECCs), to forecast stock volatility is a critical challenge that has attracted both academics and investors. While previous studies have used multimodal deep learning-based models to obtain a…","strategy":null},"assumptions":{"status":"available","items":[{"assumption":"The temporal train‑test split assumes that using only past data to predict future risk yields a valid evaluation.","type":"data","stated":"implied","quote":"We partitioned the dataset into a training set and a test set with an 8:2 ratio, organized temporally to ensure that the data in the training set precedes those in the test set."},{"assumption":"Audio recordings are accurately aligned with their text transcripts.","type":"data","stated":"implied","quote":"The dataset consists of 572 unique instances where the audio recordings were accurately and closely aligned with the text transcripts."}]},"datasets":{"status":"available","items":[{"name":"S&P 500 ECC dataset","provider":"public","kind":"alternative","access":"public","period":"2017 earnings calls","universe":"500 major companies listed on the S&P 500","quote":"The dataset utilized in this study is sourced from the publicly available S&P 500 ECC dataset as constructed by [28]. It includes both audio recordings and corresponding text transcripts from the 2017 earnings calls of 500 major companies listed on the S&P 500."}],"known":[{"id":"yahoo-finance","name":"Yahoo Finance","provider":"Yahoo","kind":"market-prices","access":"public","quote":"Additionally, for constructing volatility labels, we collected stock price information via the Yahoo Finance API1 and used equation 1 to calculate the labels for days 3, 7, 15, and 30.","mentions":1,"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 ECC Analyzer pipeline reduces average Mean Squared Error by 27.7% relative to the current state‑of‑the‑art model.","kind":"performance","quote":"Our pipeline demonstrates a 27.7% reduction in average Mean Squared Error (MSE) compared to the current state-of-the-art (SOTA) model."},{"finding":"The method yields substantial performance improvements in short‑term volatility forecasting (3‑day, 7‑day) over existing SOTA methods.","kind":"predictive-accuracy","quote":"The results show that our method achieves substantial performance improvements in forecasting short-term volatility (3-day, 7-day) compared to current SOTA methods."},{"finding":"For medium‑term (15‑day) and long‑term (30‑day) volatility, performance is comparable to the best existing methods.","kind":"performance","quote":"The results for medium-term volatility (15 days) and long-term volatility (30 days) are comparable to the best existing methods."}],"reported_performance":null,"out_of_sample":{"method":"holdout","headline_out_of_sample":null}},"independent_evidence":{"status":"none_found","supports":[],"contradicts":[],"qualifies":[],"compared_with":16,"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":"2026-11-14T18:10:14Z"},"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/"}