{"paper_id":"2508.14999","title":"Investment Portfolio Optimization Based on Modern Portfolio Theory and Deep Learning Models","authors":["Maciej Wysocki","Paweł Sakowski"],"paper_date":"2025-08-20","paper_type":"empirical","paper_url":"https://arxiv.org/pdf/2508.14999v1","page_url":"https://thequant.space/flowcharts/investment-portfolio-optimization-based-on-modern-portfolio/","scores":{"empirical_rigor":8,"math_complexity":7.5,"quadrant":"Holy Grail"},"proposes":{"status":"available","statement":"The paper proposes a novel deep‑learning‑based framework for estimating variance‑covariance matrices for portfolio optimization.","quote":"We propose a novel framework for variance‑covariance matrix estimation for purposes of the portfolio optimization, which is based on deep learning models.","summary":"This paper investigates an important problem of an appropriate variance-covariance matrix estimation in the Modern Portfolio Theory. We propose a novel framework for variancecovariance matrix estimation for purposes of the portfolio optimization, which is based on deep learning models. We employ…","strategy":null},"assumptions":{"status":"available","items":[{"assumption":"A fixed percentage commission of 50 basis points is applied to every asset when rebalancing.","type":"execution","stated":"explicit","quote":"we have decided to use such solution and consider all costs of changing position assuming a fixed percentage commission of 50 bps for every asset."},{"assumption":"365 trading days per year are used in all calculations.","type":"data","stated":"explicit","quote":"For every calculation we assumed 365 trading days per year."},{"assumption":"Missing stock quotes on weekends/holidays are forward‑filled with the last available price.","type":"data","stated":"implied","quote":"the missing stock quotes for these days were filled with the last available price"},{"assumption":"Median values of the Monte‑Carlo forecasts are used as point forecasts.","type":"model","stated":"explicit","quote":"In this study we decided to use median values as the forecasts, as it is a middle quantile and a commonly undertaken approach in probabilistic forecasting"},{"assumption":"Each rebalancing date the portfolio is built from the 10 stocks and 10 cryptocurrencies with the highest market capitalization at that date.","type":"data","stated":"implied","quote":"10 stocks and 10 cryptocurrencies with the highest market capitalization at a given date were selected and taken into the portfolio construction process"},{"assumption":"Longer observation windows improve deep‑learning model performance because they capture long‑term dependencies.","type":"model","stated":"explicit","quote":"longer observation windows translate into better performance of the deep learning models indicating that these methods require longer windows to be able"}]},"datasets":{"status":"available","items":[{"name":"Cryptocurrency daily close prices","provider":"Coinpaprika","kind":"market-prices","access":"public","period":"2018-01-01 to 2021-10-31","universe":"around 500 cryptocurrencies, top 10 by market cap each date","quote":"The dataset used in this study consists of daily close prices of cryptocurrencies"},{"name":"US stock daily close prices","provider":"Yahoo Finance","kind":"market-prices","access":"public","period":"2018-01-01 to 2021-10-31","universe":"around 1000 NYSE stocks, top 10 by market cap each date","quote":"The dataset used in this study consists of daily close prices of","known_id":"yahoo-finance"}],"known":[{"id":"coinmarketcap","name":"CoinMarketCap","provider":"CoinMarketCap","kind":"market-prices","access":"public","quote":"3 Data sourced from: coinmarketcap.com including 64 cryptocurrencies and 18 stocks.","mentions":1,"used_in":1},{"id":"yahoo-finance","name":"Yahoo Finance","provider":"Yahoo","kind":"market-prices","access":"public","quote":"4 Data and Parameters Setup 4.1 Assets The dataset used in this study consists of daily close prices of cryptocurrencies and US stocks downloaded respectively from Coinpaprika1 and Yahoo Finance services.","mentions":1,"used_in":1}],"sample":{"start":null,"end":null,"market":null,"asset_class":null,"frequency":null},"data_resolution":"daily"},"transaction_costs":{"status":"available","modeled":"yes","headline_net_of_costs":null,"detail":"fixed percentage commission of 50 bps per asset applied to all position changes","quote":"assuming a fixed percentage commission of 50 bps for every asset","source":"front_matter","rule":"R3","candidates":{"front_matter":"yes","evidence_card":"yes","evidence_card_quoted":true},"card_detail":"fixed percentage commission of 50 bps per asset applied to all position changes","card_quote":"assuming a fixed percentage commission of 50 bps for every asset","meaning":"costs are part of the reported headline results"},"results":{"status":"available","evaluation_type":"trading-pnl","findings":[{"finding":"LSTM‑RNN models deliver the highest information ratios and annualized returns among all variance‑covariance estimation methods.","kind":"performance","quote":"generally the best results in terms of the information ratio and annualized returns are obtained using the LSTM‑RNN models"},{"finding":"Longer observation windows (e.g., 90‑day and 120‑day) improve the profitability and information ratios of deep‑learning‑based strategies.","kind":"performance","quote":"The 90‑days window was the first setup when so many strategies provided positive returns and information rations"},{"finding":"Strategies with less frequent rebalancing (e.g., 30‑, 60‑, 90‑day) outperform those with very short rebalancing periods.","kind":"performance","quote":"Strategies with less frequent rebalancing typically perform better than these with the shortest rebalancing windows across all considered methods"},{"finding":"Deep learning models require longer historical windows to capture short‑ and long‑term dependencies in the variance‑covariance matrix.","kind":"statistical","quote":"longer observation windows translate into better performance of the deep learning models indicating that these methods require longer windows to be able"},{"finding":"Classical variance‑covariance estimators produce unsatisfactory, often negative, returns and information ratios, especially with longer rebalancing periods.","kind":"negative","quote":"All strategies based on the classical variance‑covariance matrix estimation methods provided rather similar results"}],"reported_performance":null,"out_of_sample":{"method":"walk-forward","headline_out_of_sample":null}},"independent_evidence":{"status":"available","supports":[{"paper":{"paper_id":"2601.17021","title":"Regret-Driven Portfolios: LLM-Guided Smart Clustering for Optimal Allocation","paper_date":"2026-01-16","authors":["Muhammad Abro","Hassan Jaleel"],"empirical_rigor":6,"page_url":"https://thequant.space/flowcharts/regret-driven-portfolios-llm-guided-smart-clustering-for-optimal-allocation/"},"question":"What rebalancing frequency leads to superior portfolio performance?","this_paper_finding":"Strategies with less frequent rebalancing (e.g., 30‑, 60‑, 90‑day) outperform those with very short rebalancing periods.","other_paper_finding":"71% of configurations that outperformed SPY used a quarterly rebalancing schedule, indicating medium‑term rebalancing is beneficial.","other_paper_quote":"Among all the parameter configurations that outperformed SPY, 71% employed a quarterly rebalancing schedule.","note":"Both find medium‑term rebalancing optimal: A shows 30‑,60‑,90‑day periods outperform very short intervals; B shows 71% of out‑performing configs use quarterly rebalancing.","confidence":0.859},{"paper":{"paper_id":"2504.12266","title":"Semiparametric Dynamic Copula Models for Portfolio Optimization","paper_date":"2025-04-16","authors":["Savita Pareek","Sujit K. Ghosh"],"empirical_rigor":6,"page_url":"https://thequant.space/flowcharts/semiparametric-dynamic-copula-models-for-portfolio-optimizat/"},"question":"Which rebalancing frequency leads to better portfolio performance?","this_paper_finding":"Strategies with less frequent rebalancing (e.g., 30‑, 60‑, 90‑day) outperform those with very short rebalancing periods.","other_paper_finding":"Weekly rebalancing with the copula‑based method yields higher gains than daily rebalancing, especially in volatile periods.","other_paper_quote":"The copula‑based method demonstrates superior performance in most scenarios, particularly under weekly rebalancing.","note":"Both findings show that longer intervals (weekly or 30‑, 60‑, 90‑day) outperform daily or very short rebalancing.","confidence":0.735}],"contradicts":[{"paper":{"paper_id":"2311.04946","title":"Causal Inference on Investment Constraints and Non-stationarity in Dynamic Portfolio Optimization through Reinforcement Learning","paper_date":"2023-11-08","authors":["Yasuhiro Nakayama","Tomochika Sawaki"],"empirical_rigor":6,"page_url":"https://thequant.space/flowcharts/causal-inference-on-investment-constraints-and-non-stationar/"},"question":"What is the impact of rebalancing frequency on portfolio performance (e.g., Sharpe ratio)?","this_paper_finding":"Strategies with less frequent rebalancing (e.g., 30‑, 60‑, 90‑day) outperform those with very short rebalancing periods.","other_paper_finding":"Unconstrained (no rebalancing frequency limit) yields the highest average Sharpe ratio; performance declines as the rebalancing interval is lengthened (weekly, bi‑weekly, monthly).","other_paper_quote":"As expected, the unconstrained condition exhibits the highest average Sharpe ratio, and performance gradually decreases as the interval is extended.","note":"A finds performance declines as rebalancing moves from no limit to weekly/bi‑weekly/monthly, while B finds 30‑, 60‑, 90‑day intervals outperform very short periods.","confidence":0.741},{"paper":{"paper_id":"2604.14439","title":"Multi periods mean-DCVaR optimization: a Recursive Neural Network resolution","paper_date":"2026-04-09","authors":["Jérôme Lelong","Véronique Maume-Deschamps","William Thevenot"],"empirical_rigor":6,"page_url":"https://thequant.space/flowcharts/multi-periods-mean-dcvar-optimization-a-recursive-neural-network-resolution/"},"question":"How does the frequency or number of portfolio rebalancing dates affect performance?","this_paper_finding":"Strategies with less frequent rebalancing (e.g., 30‑, 60‑, 90‑day) outperform those with very short rebalancing periods.","other_paper_finding":"Performance improves as the number of rebalancing dates N increases.","other_paper_quote":"Performance increases with the number of rebalancing dates N, as the policy can adapt more frequently to realized trajectories","note":"A finds less frequent (30‑, 60‑, 90‑day) rebalancing outperforms very short periods, while B reports performance improves as the number of rebalancing dates N increases.","confidence":0.656},{"paper":{"paper_id":"2305.12539","title":"Value-at-Risk-Based Portfolio Insurance: Performance Evaluation and Benchmarking Against CPPI in a Markov-Modulated Regime-Switching Market","paper_date":"2023-05-21","authors":["Peyman Alipour","Ali Foroush Bastani"],"empirical_rigor":4,"page_url":"https://thequant.space/flowcharts/value-at-risk-based-portfolio-insurance-performance-evaluat/"},"question":"What is the optimal rebalancing frequency for portfolio insurance/optimization strategies?","this_paper_finding":"Strategies with less frequent rebalancing (e.g., 30‑, 60‑, 90‑day) outperform those with very short rebalancing periods.","other_paper_finding":"Daily rebalancing yields the best Omega and Kappa measures across confidence levels.","other_paper_quote":"We observe that for all threshold values, daily rebalancing leads to better results.","note":"A finds daily rebalancing gives best Omega/Kappa; B finds 30‑, 60‑, 90‑day rebalancing outperforms very short periods.","confidence":0.655}],"qualifies":[],"compared_with":36,"citations":{"cited_by_in_corpus":1,"by_independent_authors":1,"how_cited":{"used_as_baseline":0,"uses_method_or_data":0,"extends":0,"discusses_findings":0,"mentions":1,"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/Maciej-13/vcov_forecast","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":"2026-11-14T17:29:47Z"},"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/"}