From News to Trends: A Financial Time Series Forecasting Framework with LLM-Driven News Sentiment Analysis and Selective State Spaces DOI
Renjie Wang, Minghui Sun, Limin Wang

et al.

Research Square (Research Square), Journal Year: 2025, Volume and Issue: unknown

Published: April 17, 2025

Abstract Stock price prediction is inherently challenging due to market volatility and the influence of external factors. Traditional forecasting methods primarily rely on historical data, limiting their ability capture sentiment embedded in financial news. To address this limitation, we propose Senti-MambaMoE, a novel model that integrates stock prices with information extracted from Specifically, fine-tune DeepSeek-based large language (LLM) for classification incorporate into our predictive framework. At core approach MambaMoE, which leverages efficiency state space models (SSMs) long-range dependencies while maintaining linear computational complexity, making it well-suited time series forecasting. Additionally, MoE mechanism improves model’s diverse behaviors by dynamically selecting specialized experts based data patterns. Experimental results demonstrate Senti-MambaMoE outperforms LSTM-based 23.7% Transformer-based 6.3%, highlighting its superior performance short-term prediction.

Language: Английский

PPTformer: A novel hybrid model for enhanced long-term time series forecasting with extreme value focus DOI
Jian Liu,

Junkang Guo,

Lei Gao

et al.

Knowledge-Based Systems, Journal Year: 2025, Volume and Issue: unknown, P. 113456 - 113456

Published: March 1, 2025

Language: Английский

Citations

0
Real-time control parameter update and stochastic tool wear monitoring framework for nonlinear micro-milling process DOI
Pengfei Ding, Zhijie Liu, Xianzhen Huang

et al.

Precision Engineering, Journal Year: 2025, Volume and Issue: unknown

Published: March 1, 2025

Language: Английский

Citations

0
Deep learning for algorithmic trading: A systematic review of predictive models and optimization strategies DOI Creative Commons
Mohiuddin Ahmed Bhuiyan, Md. Oliullah Rafi,

Gourab Nicholas Rodrigues

et al.

Array, Journal Year: 2025, Volume and Issue: unknown, P. 100390 - 100390

Published: April 1, 2025

Language: Английский

Citations

0
STGAT: Spatial–Temporal Graph Attention Neural Network for Stock Prediction DOI Creative Commons

Ruizhe Feng,

Shanshan Jiang,

Xingyu Liang

et al.

Applied Sciences, Journal Year: 2025, Volume and Issue: 15(8), P. 4315 - 4315

Published: April 14, 2025

Stock price prediction and portfolio optimization are critical research areas in financial markets, as they directly impact investment strategies risk management. Traditional statistical methods machine learning approaches have been widely applied to these tasks, but often fail fully capture the complex dynamics of markets. typically rely on unrealistic assumptions or oversimplified models, neglecting nonlinear high-dimensional characteristics market data. Additionally, deep methods, especially temporal convolution networks graph attention networks, introduced this area achieved significant improvements both stock optimization. Therefore, study proposes a Spatial–Temporal Graph Attention Network (STGAT) that integrates STL decomposition components structures model patterns asset correlations. By combining mechanisms with convolutional modules, STGAT effectively processes spatiotemporal data, enhancing accuracy predictions. Empirical experiments CSI 500 S&P datasets demonstrate outperforms other models performance. The portfolios constructed based STGAT’s predictions achieve higher returns real scenarios, which validates feasibility feature fusion for highlights advantages capturing characteristics. This not only provides robust tool also offers valuable insights future intelligent systems.

Language: Английский

Citations

0
From News to Trends: A Financial Time Series Forecasting Framework with LLM-Driven News Sentiment Analysis and Selective State Spaces DOI
Renjie Wang, Minghui Sun, Limin Wang

et al.

Research Square (Research Square), Journal Year: 2025, Volume and Issue: unknown

Published: April 17, 2025

Abstract Stock price prediction is inherently challenging due to market volatility and the influence of external factors. Traditional forecasting methods primarily rely on historical data, limiting their ability capture sentiment embedded in financial news. To address this limitation, we propose Senti-MambaMoE, a novel model that integrates stock prices with information extracted from Specifically, fine-tune DeepSeek-based large language (LLM) for classification incorporate into our predictive framework. At core approach MambaMoE, which leverages efficiency state space models (SSMs) long-range dependencies while maintaining linear computational complexity, making it well-suited time series forecasting. Additionally, MoE mechanism improves model’s diverse behaviors by dynamically selecting specialized experts based data patterns. Experimental results demonstrate Senti-MambaMoE outperforms LSTM-based 23.7% Transformer-based 6.3%, highlighting its superior performance short-term prediction.

Language: Английский

Citations

0