A Novel Framework for Agricultural Futures Price Prediction With BERT‐Based Topic Identification and Sentiment Analysis

Journal of Forecasting, Journal Year: 2025, Volume and Issue: unknown

Published: April 11, 2025

ABSTRACT In China's financial and economic system, the agricultural futures market plays an important role in guiding to self regulate providing efficient information transmission for regulators. The effective prediction of prices can assist production, monitoring operational risks arising from significant price fluctuations, enhancing predictability pertinence country's macroeconomic regulation policies. This study investigates main variety grain futures—soybean futures, taking into account complex non‐market influencing factors. Using historical data related news headlines soybean as source integrating topic identification sentiment analysis techniques, a novel framework predicting that integrates is constructed. model uses BERTopic extract texts, then FinBERT construct topic‐based features, fuses them with structured constructs LSTM multi‐feature inputs. order better short‐term features state transfer patterns time series, hidden Markov (HMM) further used states, which are deeply fused model. empirical results show fusing significantly improves forecasting accuracy all lags, works best forecasting, combination HMM exhibits performance advantages medium‐ long‐term forecasting. Compared baseline relies only on provide incremental contribution each feature calculated based PI metric close 50%. addition, deep learning–based performs than machine learning models dealing extreme external shocks such climate disasters, COVID‐19 pandemic, Russia–Ukraine conflict.

Language: Английский

---

A Hybrid Model Integrating Variational Mode Decomposition and Intelligent Optimization for Vegetable Price Prediction

Agriculture, Journal Year: 2025, Volume and Issue: 15(9), P. 919 - 919

Published: April 23, 2025

In recent years, China’s vegetable market has faced frequent and drastic price fluctuations due to factors such as supply–demand relationships climate change, which significantly affect government bodies, farmers, consumers, other participants in the industry supply chain. Traditional forecasting methods demonstrate evident limitations capturing nonlinear characteristics complex volatility patterns of series, underscoring necessity developing high-precision prediction models. This study proposes a hybrid model integrating variational mode decomposition (VMD), Fruit Fly Optimization Algorithm (FOA), gated recurrent unit (GRU). The employs VMD for multi-scale original series utilizes FOA adaptive optimization GRU’s critical parameters, effectively addressing challenges high nonlinearity agricultural forecasting. Empirical analysis conducted on daily data six major vegetables, specifically, Chinese cabbage, cucumber, beans, tomato, chili, radish, from 2014 2024 reveals that proposed outperforms traditional methods, single deep learning models, models predictive performance. Experimental results indicate substantial improvements key metrics including Mean Absolute Error (MAE), Root Square (RMSE), Coefficient Determination (R2), with R2 values consistently exceeding 99.4% achieving over 5% enhancement compared baseline GRU model. research establishes novel methodological framework analyzing while providing reliable technical support monitoring policy regulation.

Language: Английский

---

A hybrid TCN-XGBoost model for agricultural product market price forecasting

PLoS ONE, Journal Year: 2025, Volume and Issue: 20(5), P. e0322496 - e0322496

Published: May 2, 2025

Price volatility in agricultural markets is influenced by seasonality, supply-demand fluctuations, policy changes, and climate. These factors significantly impact production the broader macroeconomy. Traditional time series models, limited linear assumptions, often fail to capture nonlinear nature of price fluctuations. To address this limitation, we propose an integrated forecasting model that combines TCN XGBoost improve accuracy predictions. captures both short-term long-term dependencies using convolutional operations, while enhances its ability relationships. The uses 65,750 historical data points from rice, wheat, corn, with a sliding window technique construct features. Experimental results demonstrate TCN-XGBoost outperforms traditional models such as ARIMA (RMSE = 0.36, MAPE 8.9%) LSTM 0.34, 8.1%). It also other hybrid Transformer-XGBoost 0.23) CNN-XGBoost 0.29). Specifically, achieves RMSE 0.26 5.3%, underscoring superior performance. Moreover, shows robust performance across various market conditions, particularly during significant During dramatic movements, 0.28 6.1%, effectively capturing trends magnitudes changes. By leveraging TCN’s strength temporal feature extraction XGBoost’s capability complex relationships, offers efficient solution for prices. This has broad applicability, decision-making risk management.

Language: Английский

---

A Forecasting Approach for Wholesale Market Agricultural Product Prices Based on Combined Residual Correction

Applied Sciences, Journal Year: 2025, Volume and Issue: 15(10), P. 5575 - 5575

Published: May 16, 2025

Wholesale market prices of agricultural products, being essential to the daily lives consumers, are closely tied living standards and overall stability market. The use a single model predict nonlinear dynamic price time series often results in low accuracy due suboptimal available information. To address this issue, paper proposes combined residual correction-based prediction method. Initially, sparrow search algorithm (SSA) is used optimize penalty factors kernel parameters support vector regression (SVR) input weights hidden layer biases extreme learning machine (ELM), thereby improving convergence rate predictive these models. Subsequently, induced ordered weighted averaging (IOWA) operator applied determine weight vectors for SSA-SVR SSA-ELM models, reducing fluctuating accuracies individual models at different times. Finally, residuals generalized neural network (GRNN) forecasted using correction method that integrates based on IOWA operator, refining GRNN’s forecast outcomes. An empirical analysis was performed by comparing nine forecasting monthly pork Beijing. findings indicate SSA-SVR, SSA-GRNN, outperformed SVR, GRNN, ELM terms accuracy, respectively. This improvement attributed parameter optimization through SSA. proposed also showed superior compared confirm an effective tool predicting product can be other products with similar characteristics.

Language: Английский

---

How Does the Low-Frequency Investor Attention Impact the High-Frequency Copper Prices?

Published: Jan. 1, 2024

Language: Английский

---

Spatial-temporal analysis and trend prediction of regional crop disease based on electronic medical records

Applied Soft Computing, Journal Year: 2024, Volume and Issue: 167, P. 112423 - 112423

Published: Nov. 1, 2024

Language: Английский

---