{"paper_id":"2411.15674","title":"Quantile deep learning models for multi-step ahead time series prediction","authors":["Jimmy Cheung","Smruthi Rangarajan","Amelia Maddocks","Xizhe Chen","Rohitash Chandra"],"paper_date":"2024-11-24","paper_type":"empirical","paper_url":"https://arxiv.org/pdf/2411.15674v1","page_url":"https://thequant.space/flowcharts/quantile-deep-learning-models-for-multi-step-ahead-time-seri/","scores":{"empirical_rigor":7,"math_complexity":5.5,"quadrant":"Holy Grail"},"proposes":{"status":"available","statement":"The paper proposes a novel quantile regression deep learning framework for multi-step time series prediction.","quote":"We present a novel quantile regression deep learning framework for multi-step time series prediction.","summary":"Uncertainty quantification is crucial in time series prediction, and quantile regression offers a valuable mechanism for uncertainty quantification which is useful for extreme value forecasting. Although deep learning models have been prominent in multi-step ahead prediction, the development and…","strategy":null},"assumptions":{"status":"available","items":[{"assumption":"The median values for each time step are taken as the predicted values.","type":"model","stated":"explicit","quote":"we assume the median values for each time step (prediction horizon)to be the predicted values."}]},"datasets":{"status":"available","items":[{"name":"Bitcoin","provider":null,"kind":"market-prices","access":"public","period":null,"universe":null,"quote":"Our models are tested on two cryptocurrencies: Bitcoin and Ethereum, using daily close-price data"},{"name":"Ethereum","provider":null,"kind":"market-prices","access":"public","period":null,"universe":null,"quote":"Our models are tested on two cryptocurrencies: Bitcoin and Ethereum, using daily close-price data"},{"name":"Sunspots","provider":null,"kind":"benchmark","access":"public","period":null,"universe":null,"quote":"benchmark datasets that include Sunspots, Mackey-Glass and Lorenz."},{"name":"Mackey-Glass","provider":null,"kind":"benchmark","access":"public","period":null,"universe":null,"quote":"benchmark datasets that include Sunspots, Mackey-Glass and Lorenz."},{"name":"Lorenz","provider":null,"kind":"benchmark","access":"public","period":null,"universe":null,"quote":"benchmark datasets that include Sunspots, Mackey-Glass and Lorenz."}],"known":[],"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":"Quantile-ED-LSTM provides the best accuracy for all prediction horizons except step one in both univariate and multivariate strategies.","kind":"performance","quote":"Specifically, the Quantile-ED-LSTM model provides the best accuracy for all prediction horizons, except step one in the univariate and multivariate strategies."},{"finding":"Quantile-BD-LSTM consistently outperforms BD-LSTM across both univariate and multivariate strategies.","kind":"performance","quote":"the Quantile-BD-LSTM model consistently provides higher ac- curacy than its counterpart (BD-LSTM) across both univariate and multivariate strategies."},{"finding":"In the Ethereum dataset, univariate quantile models achieve the best RMSE accuracy and multivariate quantile models show the most robustness with low confidence intervals.","kind":"performance","quote":"the Univariate quantile models provide the best accuracy (RMSE) and the Multivariate quantile models provides the most robustness, as indicated by consistently low confidence intervals."},{"finding":"For Sunspot and Lorenz datasets, ED-LSTM yields the best mean RMSE, while quantile ED-LSTM performs best on the Mackey-Glass series.","kind":"performance","quote":"In the Sunspot and Lorenz datasets, ED-LSTM models provide the best mean RMSE accuracy and quantile ED-LSTM models exhibit the best performance for the Mackey-Glass time series."},{"finding":"Combining quantile regression with deep learning models is very effective even in volatile cryptocurrency markets.","kind":"predictive-accuracy","quote":"Our results demonstrated that the combining quantile regression with deep learning models has been very effective, even in volatile environments such as the cryptocurrency markets."}],"reported_performance":null,"out_of_sample":{"method":"holdout","headline_out_of_sample":null}},"independent_evidence":{"status":"none_found","supports":[],"contradicts":[],"qualifies":[],"compared_with":24,"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":"2026-11-14T17:52:47Z"},"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/"}