Research
On-device research index

arXiv research

A locally-built, LLM-digested index of recent arXiv papers in quant finance, geometry/topology, and statistical ML — keyword search served straight from SQLite on this machine.

168,742 papers · 148 categories

Trend · papers per month

4488131175 · Jun 202619922001200920172026
48 results for financial return forecasting

Deep neural networks forecast financial return distributions accurately.

problem Forecasting probability distributions of financial returns.
method Used 1D CNN and LSTM architectures with custom loss functions to optimize distribution parameters.
result LSTM with skewed Student's t distribution outperformed classical models in multiple evaluation metrics.

A new method models financial returns by separating sign and magnitude, improving forecasting accuracy.

problem Capturing nonlinear predictability in financial return dynamics.
method Decomposes returns into sign and magnitude components, using a joint distribution model.
result Significantly outperforms traditional linear models in forecasting U.S. stock market returns.

A3T-GCN model forecasts FTSE100 stock prices using technical indicators and financial ratios.

problem Forecasting closing stock prices of FTSE100 constituents.
method Hybrid A3T-GCN architecture using technical indicators, financial ratios, and sector correlations.
result A3T-GCN model improves prediction accuracy with annualized log-returns and shorter sequence lengths.

Network analysis improves stock return forecasting.

problem Improving stock return forecasting using network properties.
method Network analysis of stock return correlations, using individual and global properties of stocks.
result 50% improvement in R2 score for long-term stock returns forecasting, 3% for short-term.

Improved Hawkes model forecasts extreme financial returns more accurately.

problem Forecasting extreme tail events in financial log-returns.
method 2T-POT Hawkes model with multiple exceedance thresholds.
result 2T-POT Hawkes model outperforms GARCH-EVT model in risk forecasting.

The paper introduces a new method for forecasting financial risk using quantile-based modeling.

problem Forecasting Value-at-Risk (VaR) and Expected Shortfall (ES) for financial returns.
method Semiparametric approach using restricted quantile regression to model the conditional scale of financial returns.
result The method provides robust, distribution-free estimates of extreme losses and captures risk dynamics.

This paper fine-tunes LLMs for stock return prediction using financial news.

problem Improving stock return forecasting accuracy using LLMs.
method Fine-tuning LLMs with text and forecasting modules, comparing encoder-only and decoder-only models, and integrating token-level representations.
result LLMs' aggregated token-level embeddings enhance return predictions for long-only and long-short portfolios.

HANET combines LSTM and attention mechanisms for better financial forecasting.

problem Lack of distinct macroeconomic regimes in financial datasets.
method Hierarchical Cross-Attention mechanism integrating long-run macro contexts with high-frequency market dynamics.
result HANET outperforms neural forecasters, especially during turbulent periods.

CSHT predicts financial returns from news using a novel transformer model on a sphere.

problem Financial forecasting from news and sentiment.
method Granger-causal hypergraph structure, Riemannian geometry, causally masked Transformer attention.
result CSHT outperforms baselines in return prediction, regime classification, and asset ranking.

The study compares differencing methods for financial data and finds fractional differencing improves model performance.

problem Improving financial time series forecasting models using appropriate data transformation techniques.
method Comparative analysis of traditional logarithmic returns and fractional differencing methods, including tempered extensions.
result Fractional differencing methods improve model forecasting performance and trading strategy effectiveness.

Diffolio uses a diffusion model for multivariate financial forecasting and portfolio construction.

problem Probabilistic forecasting of multivariate financial time-series with complex cross-sectional dependencies.
method Diffolio employs a denoising network with hierarchical attention architecture, incorporating asset-level and market-level layers and a correlation-guided regularizer.
result Diffolio outperforms various probabilistic forecasting baselines in multivariate forecasting accuracy and portfolio performance.

Enhanced financial forecasting using supervised autoencoders with noise augmentation and triple labeling.

problem Improving investment strategy performance on noisy financial data.
method Supervised autoencoders with noise augmentation and triple barrier labeling.
result Supervised autoencoders with balanced noise augmentation and bottleneck size significantly boost strategy effectiveness.

Enhanced financial forecasting with supervised autoencoders for S&P 500 and cryptocurrencies.

problem Improving investment strategy performance in financial markets.
method Supervised autoencoders with noise augmentation and triple barrier labeling.
result Supervised autoencoders with balanced parameters significantly boost strategy effectiveness.

ChatGPT predicts stock market reactions from news headlines without financial training.

problem Predicting stock price movements using non-financial data.
method Used post-knowledge-cutoff headlines to train ChatGPT-4, which forecasts stock market reactions.
result ChatGPT-4 can predict stock market reactions with high accuracy, especially for small stocks and negative news.

DBNs improve VaR forecasting compared to traditional models, but SVaR forecasts are conservative.

problem Forecasting VaR and SVaR using dynamic Bayesian networks.
method DBN framework applied to S&P 500 index returns, comparing to autoregressive models and historical simulation.
result DBNs achieve comparable VaR forecasting accuracy to historical simulation models, but SVaR forecasts remain conservative.

The study examines how posterior drift affects forecasting accuracy in overparametrized models, particularly in financial markets.

problem Impact of posterior drift on out-of-sample forecasting accuracy in overparametrized models.
method Investigation of posterior drift and its effect on model performance in financial markets.
result Overparametrized models can be sensitive to sub-periods and bandwidth parameters, leading to inconsistent returns.

Volatility forecasting and return prediction in high-frequency Chinese equity markets.

problem Improving statistical forecasting performance and economic strategy outcomes in equity markets.
method Developing a sequential two-stage framework combining realized volatility modeling and XGBoost return prediction.
result Regime-aware volatility forecasting outperforms baseline models.

Neural Lévy model improves risk and density forecasting for financial returns.

problem Financial returns exhibit heavy tails, volatility clustering, and jumps.
method Proposes a neural Lévy jump-diffusion framework that learns conditional drift, diffusion, jump intensity, and size distribution.
result Demonstrates improved calibration, sharper tail control, and risk reduction.

BOA improves financial forecasting by combining expert models.

problem Challenges in choosing between multiple machine learning models for financial forecasting.
method Online aggregation of expert models using Bernstein Online Aggregation (BOA) procedure.
result BOA leads to better portfolio performance, higher Sharpe Ratio, and lower shortfall.

Paper proposes a joint quantile regression for VaR and ES forecasting.

problem Forecasting Value at Risk (VaR) and Expected Shortfall (ES) of multiple assets simultaneously.
method Multivariate quantile regression framework with time-varying process for VaR and ES.
result The proposed method outperforms other models in risk measure forecasts.

Study improves cryptocurrency price prediction using unlabeled text data.

problem Predicting cryptocurrency returns from unlabelled text data.
method Introduced weak learning approach to finetune BERT on unlabeled text data.
result Finetuning pretrained NLP models with weak labels enhances forecast accuracy.

Foundation models improve volatility forecasting in finance.

problem Improving volatility forecasting in financial markets.
method Evaluation of TimesFM model, incremental fine-tuning, comparison with econometric benchmarks.
result Incremental fine-tuning improves forecast accuracy and outperforms traditional models.

Adaptive volatility method improves probabilistic financial forecasting.

problem Probabilistic forecasting in financial markets.
method Adapts classical time-varying volatility models with online stochastic optimization.
result Ranked 5th in M6 financial forecasting competition.

A new model forecasts Value-at-Risk using NIG distribution and dynamic scores.

problem Forecasting Value-at-Risk (VaR) in financial markets.
method Proposes a parametric forecasting model based on the normal inverse Gaussian distribution (NIG) incorporating intraday information.
result The model outperforms traditional GARCH models, especially in high-risk scenarios.

Three adaptive methods improve financial forecasting and portfolio management.

problem Improving financial forecasting and portfolio management in volatile markets.
method Dynamic Model Selection (DMS), Adaptive Ensemble (AE), Dynamic Asset Allocation (DAA).
result Adaptive methods outperform long-only benchmarks in US market returns.

Machine learning models outperform traditional CAPM in forecasting financial asset prices.

problem Predicting and forecasting financial asset prices and returns.
method Comparison of modern Machine Learning algorithms with the Capital Asset Pricing Model (CAPM) on U.S. equities data.
result Implemented Machine Learning models significantly outperform the CAPM on out-of-sample test data.

Long short-term memory network outperforms seasonal model in JSE Top 40 forecasting.

problem Comparing neural network performance to traditional models in financial forecasting.
method Used long short-term memory network for JSE Top 40 return data forecasting.
result Long short-term memory network outperforms seasonal model in forecasting.

This paper forecasts cryptocurrency log-returns using LASSO-VAR and sentiment analysis.

problem Forecasting log-returns of cryptocurrencies using social media sentiment.
method LASSO-VAR model combined with Twitter and Reddit sentiment data.
result The model predicts the correct direction of cryptocurrency returns more than 50% of the time.

TSFMs improve financial forecasting from diverse datasets.

problem Challenges in forecasting financial time series due to noisy, non-stationary, and heterogeneous data.
method Empirical study of TSFMs in global financial markets, evaluating zero-shot inference, fine-tuning, and pre-training from scratch.
result Pre-trained TSFMs on financial data achieve substantial forecasting and economic improvements, highlighting the value of domain-specific adaptation.

Transformer-based models overfit financial time series data, leading to increased prediction variance.

problem Forecast collapse of transformer-based models under squared loss in financial time series.
method Theoretical analysis and numerical experiments on high-frequency EUR/USD exchange rate data.
result Increased model expressivity in Transformer-based models leads to spurious fluctuations without reducing bias, resulting in higher prediction variance.

This paper evaluates various loss functions for Transformer models in stock ranking.

problem Evaluating loss functions for Transformer models in stock ranking.
method Systematic evaluation of advanced loss functions (pointwise, pairwise, listwise) on S&P 500 data.
result Different loss functions impact a model's ability to discern profitable relative orderings among assets.

Model forecasts market structure from financial networks using machine learning.

problem Predicting market correlation structure from financial networks.
method Dynamic Asset Graph (DAG), Dynamic Minimal Spanning Tree (DMST), Dynamic Threshold Networks (DTN).
result Model improves market structure forecasting by up to 40% over benchmarks.