{"paper_id":"2502.10776","title":"A Distillation-based Future-aware Graph Neural Network for Stock Trend Prediction","authors":["Zhipeng Liu","Peibo Duan","Mingyang Geng","Bin Zhang"],"paper_date":"2025-02-15","paper_type":"empirical","paper_url":"https://arxiv.org/pdf/2502.10776v1","page_url":"https://thequant.space/flowcharts/a-distillation-based-future-aware-graph-neural-network-for-s/","scores":{"empirical_rigor":5.5,"math_complexity":7.5,"quadrant":"Holy Grail"},"proposes":{"status":"available","statement":"The paper proposes a novel distillation-based future-aware graph neural network framework (DishFT-GNN) that trains a teacher and a student model iteratively to capture historical-future distribution correlations for stock trend prediction.","quote":"we propose a novel distillation-based future-aware GNN framework (DishFT-GNN) for stock trend prediction.","summary":"Stock trend prediction involves forecasting the future price movements by analyzing historical data and various market indicators. With the advancement of machine learning, graph neural networks (GNNs) have been extensively employed in stock prediction due to their powerful capability to capture…","strategy":null},"assumptions":{"status":"available","items":[{"assumption":"A stock’s future dynamics can be revealed from its historical patterns.","type":"model","stated":"implied","quote":"Typically, DL-based stock trading methods are based on the fundamental hypothesis that a stock’s future dynamics can be revealed from its historical patterns"},{"assumption":"Stock trend prediction is modeled as a binary node classification task.","type":"data","stated":"explicit","quote":"We formulate stock trend prediction as a binary node classification task."}]},"datasets":{"status":"available","items":[{"name":"S&P 100","provider":null,"kind":"market-prices","access":"public","period":"January 1, 2019, to September 30, 2023","universe":"96 stocks selected from S&P 100","quote":"we conduct extensive experiments using two datasets from American stock indices, i.e., S&P 100 and NASDAQ 100, spanning from January 1, 2019, to September 30, 2023."},{"name":"NASDAQ 100","provider":null,"kind":"market-prices","access":"public","period":"January 1, 2019, to September 30, 2023","universe":"94 stocks selected from NASDAQ 100","quote":"we conduct extensive experiments using two datasets from American stock indices, i.e., S&P 100 and NASDAQ 100, spanning from January 1, 2019, to September 30, 2023."}],"known":[],"sample":{"start":null,"end":null,"market":null,"asset_class":null,"frequency":null},"data_resolution":"daily"},"transaction_costs":{"status":"available","modeled":"unclear","headline_net_of_costs":null,"detail":null,"quote":null,"source":"conflict","rule":"R4b","candidates":{"front_matter":"n/a","evidence_card":"no","evidence_card_quoted":true},"conflict_kind":"applicability","card_detail":null,"card_quote":"we use the classic GCN as the backbone and compare the investment profits with and without DishFT-GNN for back-testing, which is illustrated in Fig.3.","meaning":"the platform classifier and the paper-text extraction disagree, so no value is asserted"},"results":{"status":"available","evaluation_type":"forecast-accuracy","findings":[{"finding":"DishFT-GNN improves accuracy by over 2% in most cases and up to 5.41% when integrated with MGAR, and boosts MCC by 82.3% to 218% across the two datasets.","kind":"performance","quote":"it achieves an accuracy improvement over 2% in most cases and up to 5.41% when integrating MGAR, and an MCC improvement of 82.3% to 218% across two datasets."},{"finding":"The profit gap between the baseline GCN and the GCN enhanced with DishFT-GNN widens over time in back-testing, indicating higher profitability.","kind":"performance","quote":"It is obvious that the profit gap between two methods gradually widens over time, which further verify the effectiveness of DishFT-GNN."}],"reported_performance":null,"out_of_sample":{"method":"holdout","headline_out_of_sample":null}},"independent_evidence":{"status":"none_found","supports":[],"contradicts":[],"qualifies":[],"compared_with":37,"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":true,"code_url":"https://github.com/ZhupengLiu/DisFT-GNN","note":null},"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":"2026-11-14T17:45:30Z"},"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/"}