{"paper_id":"2309.02072","title":"Global Neural Networks and The Data Scaling Effect in Financial Time Series Forecasting","authors":["Chen Liu","Minh-Ngoc Tran","Chao Wang","Richard Gerlach","Robert Kohn"],"paper_date":"2023-09-05","paper_type":"empirical","paper_url":"https://arxiv.org/pdf/2309.02072v6","page_url":"https://thequant.space/flowcharts/global-neural-networks-and-the-data-scaling-effect-in-financ/","scores":{"empirical_rigor":8,"math_complexity":6.5,"quadrant":"Holy Grail"},"proposes":{"status":"available","statement":"The paper proposes a global training approach for volatility forecasting that pools information across thousands of stocks.","quote":"By focusing on volatility forecasting, we employ a dataset comprising over 10,000 global stocks and implement a global estimation strategy that pools information across cross-sections.","summary":"Neural networks have revolutionized many empirical fields, yet their application to financial time series forecasting remains controversial. In this study, we demonstrate that the conventional practice of estimating models locally in data-scarce environments may underlie the mixed empirical…","strategy":null},"assumptions":{"status":"available","items":[{"assumption":"Conditional return variance follows a Gaussian error distribution.","type":"model","stated":"explicit","quote":"In volatility modeling, the key quantity of interest is the conditional return variance, σ2 t = var(yt | Ft−1), where Ft−1 contains a set of available information up to time t −1, which, in our case, is the past returns. A local volatility model, assuming Gaussian errors, is specified as yt = σtϵt, ϵt i.i.d ∼N(0, 1),"},{"assumption":"All series share a common conditional volatility function f* in the global model.","type":"model","stated":"explicit","quote":"Global training uses several time series to estimate a shared set of parameters, forming a global model."},{"assumption":"Stocks with fewer than 1,260 observations are excluded from training.","type":"data","stated":"explicit","quote":"To ensure the local models can be effectively trained, stocks with fewer than 1,260 (five years) observations during the training period are excluded, leaving a total of 11,771 stocks."},{"assumption":"Training, validation, and testing periods are non‑overlapping calendar years.","type":"data","stated":"explicit","quote":"period spans from 2014 to 2019, the validation period spans from 2020 to 2021, and the testing period spans from 2022 to 2023."},{"assumption":"When evaluating supervised forecasting for the data‑scaling study, only the 10 smallest‑training‑set stocks are used.","type":"execution","stated":"explicit","quote":"For supervised forecasting, evaluation is restricted to the 10 stock series in the smallest training set rather than all 10,240 training series."}]},"datasets":{"status":"available","items":[{"name":"Global stock dataset","provider":null,"kind":"market-prices","access":"unknown","period":"2014-2023","universe":"11,771 global stocks","quote":"Of the 11,771 stock series in our dataset, 10,240 are allocated for training and"}],"known":[{"id":"refinitiv","name":"Refinitiv / LSEG (Datastream, Eikon, Tick History)","provider":"LSEG","kind":"market-prices","access":"commercial","quote":"2.3 Dataset We collected the daily closing prices of more than 12000 stocks from 10 exchanges through the Reuters Refinitiv Workspace, covering the decade from January 1, 2014, to January 1, 2024.","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":"Forecasting accuracy improves markedly as the training dataset becomes larger and more heterogeneous.","kind":"predictive-accuracy","quote":"Our econometric analysis reveals that forecasting accuracy improves markedly as the training dataset becomes larger and more heterogeneous."},{"finding":"Even with as little as 12 months of data, globally trained networks deliver robust predictions for individual stocks and portfolios not in the training set.","kind":"predictive-accuracy","quote":"Notably, even with as little as 12 months of data, globally trained networks deliver robust predictions for individual stocks and portfolios that are not even in the training dataset."},{"finding":"Global models trained on a similarly sized but more diverse dataset generally outperform group‑specific models.","kind":"performance","quote":"The results indicate that the global model, trained on a similarly sized but more diverse dataset, generally outperforms the group-specific models."},{"finding":"The global training approach and adequate data size are essential for the success of neural‑network models in finance.","kind":"theoretical","quote":"Our findings show that the global training approach and adequate data size are essential for the success of NN models."},{"finding":"When employing flexible models such as neural networks, stock series should be effectively treated as homogeneous and modeled together for optimal performance.","kind":"theoretical","quote":"Moreover, the results again highlight that, when employing flexible models such as neural networks, stock series should be effectively treated as homogeneous and modeled together for optimal performance."}],"reported_performance":null,"out_of_sample":{"method":"holdout","headline_out_of_sample":null}},"independent_evidence":{"status":"available","supports":[{"paper":{"paper_id":"2302.01668","title":"Empirical analysis in limit order book modeling for Nikkei 225 Stocks with Cox-type intensities","paper_date":"2023-02-03","authors":["Shunya Chomei"],"empirical_rigor":7.5,"page_url":"https://thequant.space/flowcharts/empirical-analysis-in-limit-order-book-modeling-for-nikkei-225-stocks-with-cox/"},"question":"Does using more frequent or larger training data improve forecasting accuracy of limit‑order‑book models?","this_paper_finding":"Forecasting accuracy improves markedly as the training dataset becomes larger and more heterogeneous.","other_paper_finding":"Calibration every 1‑2 weeks (instead of daily) can increase accuracy from ~77% to ~78%, indicating benefit from more training data.","other_paper_quote":"Also, in the “imb n e es la 1 (n = 1, 2)” model, the accuracy can be improved from about 77% to about 78% by calibrating every 1‑2 weeks instead of daily.","note":"Both findings show accuracy gains with more data: A’s 1‑2‑week calibration raises accuracy from ~77% to ~78%; B reports marked improvement as training set grows and diversifies, despite different markets.","confidence":0.832},{"paper":{"paper_id":"2309.00875","title":"A hidden Markov model for statistical arbitrage in international crude oil futures markets","paper_date":"2023-09-02","authors":["Viviana Fanelli","Claudio Fontana","Francesco Rotondi"],"empirical_rigor":7.5,"page_url":"https://thequant.space/flowcharts/a-hidden-markov-model-for-statistical-arbitrage-in-international-crude-oil/"},"question":"How does training sample size affect model performance in statistical arbitrage and neural‑network forecasting?","this_paper_finding":"Forecasting accuracy improves markedly as the training dataset becomes larger and more heterogeneous.","other_paper_finding":"A minimum sample size is required for accurate estimation; smaller training samples weaken cointegration and reduce performance.","other_paper_quote":"reducing the training sample weakens the cointegrating relationship among the time series and, as a consequence, lowers the overall performance of our strategies","note":"Both agree larger training sets improve performance; A warns a minimum size is needed to preserve cointegration, B shows accuracy rises with bigger, more heterogeneous data.","confidence":0.778},{"paper":{"paper_id":"2501.01278","title":"Risk forecasting using Long Short-Term Memory Mixture Density Networks","paper_date":"2025-01-02","authors":["Nico Herrig"],"empirical_rigor":6,"page_url":"https://thequant.space/flowcharts/risk-forecasting-using-long-short-term-memory-mixture-densit/"},"question":"Does increasing the training dataset size improve forecasting accuracy of neural network models?","this_paper_finding":"Forecasting accuracy improves markedly as the training dataset becomes larger and more heterogeneous.","other_paper_finding":"Model performance is highly dependent on proper initialization and large training datasets.","other_paper_quote":"Further, results indicate that LSTM-MDNs require a large amount of training data to appropriately account and \"learn\" underlying patterns from the data.","note":"Both findings agree that larger, more heterogeneous training sets boost performance; B additionally stresses proper initialization.","confidence":0.582}],"contradicts":[],"qualifies":[{"paper":{"paper_id":"2611.23126","title":"Which Past Matters? Training History and Multi-Forecast Horizon in Equity Transformers","paper_date":"2026-11-17","authors":["Chariton Chalvatzis"],"empirical_rigor":8,"page_url":"https://thequant.space/flowcharts/which-past-matters-training-history-and-multi-forecast-horizon-in-equity/"},"question":"Does increasing the size and heterogeneity of training data improve forecasting accuracy?","this_paper_finding":"Forecasting accuracy improves markedly as the training dataset becomes larger and more heterogeneous.","other_paper_finding":"Increasing training quantity does not uniformly improve accuracy; the benefit of larger samples is conditional on the sampling family and horizon.","other_paper_quote":"Increasing training quantity does not uniformly improve accuracy while an equal-weight ensemble of all twelve models exceeds every individual mean at 15, 20 and 25 minutes, but not at the remaining horizons. Shared architecture and inputs do not produce identical forecasts. not uniformly improve accuracy and the progression toward larger samples at longer horizons is conditional on the sampling family.","note":"A finds accuracy improves markedly with larger, more heterogeneous data; B finds larger samples help only conditionally, depending on sampling family and forecast horizon.","confidence":0.621}],"compared_with":35,"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":"2026-11-14T18:22:30Z"},"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/"}