Stock trend prediction plays a central role in investment decision-making and has attracted extensive attention from both investors and institutions. Although recent studies have employed graph structures to model the complex relationships among financial entities, existing models fail to efficiently capture semantically rich edge features between heterogeneous entities, thereby limiting the ability to fuse and align multimodal data such as market indicators, financial events, and heterogeneous graph structures. Therefore, we propose a multimodal edge-enhanced heterogeneous graph transformer with LLM-driven knowledge graphs (MEHGT-LKG) for stock trend prediction. Specifically, we first fine-tune a large language model (LLM) by using instruction tuning datasets to design a financial event-centric knowledge extraction agent (FinEX). Subsequently, we encode the structured tuples generated from FinEX into financial event-centric knowledge graphs (FEKGs) and construct multimodal heterogeneous graphs by incorporating multimodal information. Finally, we design a multimodal edge-enhanced heterogeneous graph transformer (MEHGT) to fully encode a series of semantically enriched multimodal heterogeneous graphs spanning different time horizons. MEHGT models edge-level features through type-specific encoders and integrates them into both multi-head attention and message passing, significantly enriching the representation of relational semantics and target nodes. Extensive experimental results on multiple datasets demonstrate that the proposed method outperforms the state-of-the-art methods.

Multimodal Emotion Recognition in Conversation (MERC) aims to identify the emotional state of a speaker who expresses their opinions through text, vision, and audio information during conversations. MERC enables intelligent machines to exhibit empathy, which can increase the effectiveness of human–computer interactions. However, the existing research lacks sufficient mining of the emotional and semantic information of multiple modalities, as well as the differences and associations between multiple modalities. Hence, the two core focuses of this study are the multimodal feature mining method and the emotion fusion method in the conversational context. We propose a Feature-Enhanced Multimodal Interactive (FEMI) model for MERC tasks. Specifically, the proposed FEMI model is designed considering the following three objectives: (1) designing a feature-enhanced module that contains different feature extractors to explore deep emotional and semantic information from emotional clues and semantic attributes; (2) building a dialogue incremental transformer module to reconstruct the context interaction between interlocutors; and (3) proposing a multimodal interactive module to eliminate multimodal differences and build multimodal and cross-modal emotional associations. Extensive experiments were performed on two public datasets, and the results demonstrated that the proposed FEMI model is superior to MERC tasks.

The strategic choice of model "openness" has become a defining issue for the foundation model (FM) ecosystem. While this choice is intensely debated, its underlying economic drivers remain underexplored. We construct a two-period game-theoretic model to analyze how openness shapes competition in an AI value chain, featuring an incumbent developer, a downstream deployer, and an entrant developer. Openness exerts a dual effect: it amplifies knowledge spillovers to the entrant, but it also enhances the incumbent's advantage through a "data flywheel effect," whereby greater user engagement today further lowers the deployer's future fine-tuning cost. Our analysis reveals that the incumbent's optimal first-period openness is surprisingly non-monotonic in the strength of the data flywheel effect. When the data flywheel effect is either weak or very strong, the incumbent prefers a higher level of openness; however, for an intermediate range, it strategically restricts openness to impair the entrant's learning. This dynamic gives rise to an "openness trap," a critical policy paradox where transparency mandates can backfire by removing firms' strategic flexibility, reducing investment, and lowering welfare. We extend the model to show that other common interventions can be similarly ineffective. Vertical integration, for instance, only benefits the ecosystem when the data flywheel effect is strong enough to overcome the loss of a potentially more efficient competitor. Likewise, government subsidies intended to spur adoption can be captured entirely by the incumbent through strategic price and openness adjustments, leaving the rest of the value chain worse off. By modeling the developer's strategic response to competitive and regulatory pressures, we provide a robust framework for analyzing competition and designing effective policy in the complex and rapidly evolving FM ecosystem.

Accurate prediction of the stock indexes in the new energy market is of significant importance to both investors and policymakers. However, in response to the volatility and uncertainty characteristic of the new energy market, most scholars currently focus on training prediction methods using features from a single time scale, which cannot capture the fluctuations of new energy stock indexes under different time scales. Therefore, in this paper, a hybrid deep learning model Multi-kernel Parallel TemporalNet (MKP-TemporalNet) is proposed for predicting new energy stock indexes. This model initially incorporates an attention mechanism to calibrate the feature importance of multivariate time series dynamically, and then combines the characteristics of improved Temporal Convolutional Networks (iTCN) and Bidirectional Gated Recurrent Units (BiGRU) to enhance prediction accuracy effectively. In particular, the novelty of the iTCN lies in the development of a multi-kernel parallel convolution structure within a residual layout at the core of the temporal convolution module, to address the low efficiency of traditional TCN's single kernel convolution in extracting temporal features from input sequences at different time scales. Results from evaluating MKP-TemporalNet against several popular machine learning models on six new energy stock indexes confirm its predictive effectiveness in the new energy sector.

This paper focuses on boundary disturbance rejection control (BDRC) for fractional-order multi-agent systems (FMASs) with reaction diffusion terms. Firstly, based on the output information, the corresponding observer is designed to obtain information about all the agents and the unknown external perturbations. Then, a novel distributed boundary control protocol is designed based on the observed values. Thus, the convergence analysis of the closed-loop system is accomplished through the Mittag-Leffler stability theory and the linear matrix inequality condition. Finally, a simulation example is used to verify the correctness of the results.

Accurate interval-valued stock price prediction is challenging and of great interest to investors and for-profit organizations. In this study, by considering individual stock information and relevant stock information simultaneously, we propose a novel interval dual convolutional neural network (Dual-CNNI) model based method to predict interval-valued stock prices. First, the individual and relevant stock information are collected and transformed into images. Then, the Dual-CNNI model is proposed to predict interval-valued stock prices. Specifically, two convolutional neural network (CNN) models with different structures are constructed to respectively extract individual stock features and relevant stock features, and then an interval multilayer perceptron (MLPI) model is used for final interval-valued stock price prediction. Finally, extensive experiments are conducted based on six randomly selected stocks, with comparison to several popular machine learning model based methods and interval-valued time series (ITS) prediction methods. The experimental results indicate that the proposed Dual-CNNI based method has superior predictive ability.

Carbon emissions trading is pivotal for advancing China's low-carbon goals. As the primary tradable asset in the carbon market, carbon emission allowances inevitably experience price fluctuations. However, numerous empirical studies show that the frequency of real-world data is highly unstable, which results in the failure of probabilistic modeling. Therefore, this paper aims to model the dynamics of carbon emission allowance prices in China using four mainstream uncertain differential equations. The optimal model is chosen through rolling window cross-validation using the criterion of minimizing average testing errors. Parameters of the optimal model are determined by moment estimation based on residuals, and the model's effectiveness is also assessed through uncertain two-sided hypothesis testing. Additionally, we forecast carbon emission allowance prices and their 95% confidence intervals for the next 14 business days. To manage trading risks, we propose a customized carbon option contract for pricing European carbon options and conduct sensitivity analysis on key parameters. Finally, we present a paradox of stochastic differential equations for modeling carbon emission allowance prices.

Waste incineration technology has received extensive attention for its advantages of being harmless, reducing, and recycling. However, the waste-to-energy incineration project confronts significant “not-in-my-backyard (NIMBY) concerns,” and irrational location choices will have negative effects on the project’s economy and sustainability; it is also a great challenge to the credibility of the government. To this end, a multi-criteria decision-making framework is constructed for the site selection of waste-to-energy incineration projects. To begin with, a site selection criteria system is established including 16 sub-criteria from four aspects, where probabilistic linguistic term sets are introduced to depict the qualitative sub-criteria and probabilistic hesitant fuzzy sets are employed to express the uncertainty of quantitative sub-criteria. An optimization model is then built to determine the weights of criteria based on the Pearson correlation coefficient and least square method. Furthermore, a regret-preference ranking organization methods for enrichment evaluations (PROMETHEE) model is presented to rank alternatives in a heterogeneous decision environment. Finally, a case study in China is conducted to validate the applicability of the proposed framework; the result of the site selection demonstrates that alternative A2 located in Miyun, Beijing, is the most optimal option. This work provides investors with scientific decision reference and also extends the methods in the decision-making field.

As the second largest bond market in the world, China's bond market has attracted extensive attention in recent years. Given its importance in facilitating financing arrangements and informing investment decisions, accurate bond coupon prediction is valuable. This paper proposes an ensemble model combining TabNet, DeepFM, and XGBoost for predicting the coupons of investment-grade corporate bonds. Specifically, to optimize the hyperparameters of the proposed model, an improved butterfly optimization algorithm incorporating the concepts of good point sets, refraction opposition-based learning, switching probability adjustment, and Solis & Wets search strategies is developed. Extensive experiments using data on China's investment-grade corporate bonds demonstrate the superior performance of the proposed model in the accuracy of bond coupon predictions. Additionally, the importance of various features has been discussed. The results show that the base interest rate for valuation and term to maturity are important to bond coupon predictions obtained by the proposed model.

Stock price prediction with financial news is beneficial for making correct investment decisions. Recent researches mainly focus on extracting sentiments from news. The publishers as a crucial part of news, which can reflect the authority and reliability, but are usually neglected. In this paper, we propose a novel hybrid deep learning model for stock price prediction that considers the financial news publishers. Specifically, a novel feature about financial news publishers is constructed by classifying financial news publishers. Furthermore, a novel hybrid model named 2d-CNN-AM-LSTM is proposed by combining Convolutional Neural Network (CNN), Attention Mechanism (AM) and Long Short-term Memory, so that it can fuse textual and numerical data in-depth. To verify the superiority of the proposed model, numerous baselines are employed to conduct experiments with representative stock datasets. The experimental analysis demonstrates that the proposed model considering financial news publishers surpasses other benchmarks and achieves outstanding prediction performance.

Compared with point data, interval data can better grasp the internal structural characteristics of financial markets from a global perspective. However, the existing prediction research on interval data only focuses on the single prediction of error series or the preprocessing of original series, and the methods usually cannot fully extract the main features of interval stock price time series. Therefore, this paper proposes an error correction and decomposition method for forecast of interval-valued stock price time series. In view of the role of the error series in the combination forecasting model, we first use Ljung-Box test and machine learning model to test and modify the interval-valued error series generated by the original series. Then,bivariate empirical mode decomposition technique is used to decompose the corrected error series into multiple intrinsic mode functions（IMFs） and a residual. Then, a single machine learning model is used to predict each IMFs component and residual except IMF1 component. Finally,the predicted values of the original series and the error series are aggregated to reconstruct the predicted values of the interval stock price. Furthermore, on the basis of the proposed method, we build an interval-valued stock price forecasting model based on error correction and decomposition, and use real stock market data for empirical analysis. Experimental results show that the proposed method is superior to some traditional methods in prediction accuracy.

With the opening of the Stock Connect programs, the mainland China and Hong Kong stock markets are becoming more closely linked. In this paper, we develop a China's stock market risk early warning system. The proposed early warning system consists of three components. First, we use value at risk (VaR) to identify the stock market risk in which stock market risk is divided into multiple categories instead of two categories. Second, we construct a comprehensive indicator system in which basic indicators, technical indicators, overseas return rate indicators, and macroeconomic indicators are considered simultaneously. Third, we use four machine learning models, namely long short-term memory (LSTM), gate recurrent unit (GRU), multilayer perceptron (MLP), and EXtreme Gradient Boosting algorithm (XGBoost), to predict China's stock market risk. Experimental results show that: (1) Considering the macroeconomic indicators and basic indicators of Shanghai Composite Index (SSEC), ShenZhen Component Index (SZCZ) and Hang Seng Index (HSI) can significantly improve the performance of predicting China's stock market risk. (2) The opening of SH-HK Stock Connect program improves the predictive performance, but the opening of SZ-HK Stock Connect program decreases the predictive performance. (3) The indicators related to Hong Kong become more important after the SZ-HK Stock Connect program.

The water-energy crisis seriously affects the sustainable development of China's steel industry chain. To achieve a coordinated development new path from the perspective of circular economy, it is necessary to analyze “water-energy-economy” dependency relationship of the steel products. This study analyzes a variety of steel products from the perspective of industry chain and simulates the “water-energy-economy” potential changes of products under different scenarios by developing a multi-objective optimization model. In this model, Random Forest (RF) and GEne Network Inference with Ensemble of trees (GEINIE3) algorithms are combined to evaluate the 2013–2019 “water-energy-economy” dependency relationships firstly. Then, improved Quantum Particle Swarm Optimization (QPSO) algorithm is applied to dynamically simulate potential changes of water, energy and economic performance in the steel industry chain under different scenarios, and to design an optimal development path from the perspective of optimizing economic performance within minimum water and minimum energy use constraints. Results firstly point out the current “water-energy-economy” triple dimension dependency relationship of China's steel industry is weak. Secondly, the “water-economy” dependence has changed from one-way dependence to two-way dependence, and the “energy-economy” relationship still shows a one-way dependence. Finally, when improving resource utilization rate, assigning priority to the reuse of scrap steel, while restricting pig iron and primary steel use, may help maximize the coordinated development of “water-energy-economy” in the steel industry chain. Policy implications are proposed based on the results and provided decision-making basis for the country and relevant enterprises to promote sustainable development of the steel industry chain.

Relationship prediction in knowledge graph aims to identify and infer new relationships from existing data, and provides knowledge services for many downstream tasks. At present, many researches solve the link prediction problem between entities by mapping entities and relations into a vector space or searching the paths between entities. These methods only consider the influence of single path or first-order information but ignore more complex relation information between entities. Therefore, this paper proposes a novel link prediction method based on subgraph reasoning in knowledge graph, uses the subgraph structure to obtain the entity pair neighborhood structure information, combines the advantages of representation learning and path reasoning, and realizes the relationship prediction between entities. This paper first extends the paths between entities to subgraphs, constructs node subgraph and relationship subgraph from entity level and relationship level respectively, then combines the graph embedding representation with the graph neural network to calculate the subgraph features, to get richer entity characteristics and relationship characteristics. Finally, this paper calculates the neighborhood structure information of entity pairs from the subgraph structure to conduct link prediction between entities. Experimental results demons-trate that the proposed approach outperforms other reasoning-based link prediction methods on two benchmark datasets.

The performance of portfolio model can be improved by introducing stock prediction based on machine learning methods. However, the prediction error is inevitable, which may bring losses to investors. To limit the losses, a common strategy is diversification, which involves buying low-correlation stocks and spreading the funds across different assets. In this paper, a diversified portfolio selection method based on stock prediction is proposed, which includes two stages. To be specific, the purpose of the first stage is to select diversified stocks with high predicted returns, where the returns are predicted by machine learning methods, i.e. random forest (RF), support vector regression (SVR), long short-term memory networks (LSTM), extreme learning machine (ELM) and back propagation neural network (BPNN), and the diversification level is measured by Pearson correlation coefficient. In the second stage, the predictive results are incorporated into a modified mean-variance (MMV) model to determine the proportion of each asset. Using China Securities 100 Index component stocks as study sample, the empirical results demonstrate that the RF+MMV model achieves better results than similar counterparts and market index in terms of return and return-risk metrics.

In this paper, we identify the dynamic influence of financial institutions based on a complex network modelling method. We first construct a financial network based on stock comprehensive evaluation (SCE), which is obtained by the technique for order preference by similarity to an ideal solution (TOPSIS). Then, the dynamic influence of financial institutions is identified by iterating the static influence of each network. The results indicate that (i) the dynamic influence of financial institutions is greater than their static influence in analysing the evolution of influence and (ii) banks and securities institutions play an important role in the financial system.

The success of portfolio construction depends primarily on the future performance of stock markets. Recent developments in machine learning have brought significant opportunities to incorporate prediction theory into portfolio selection. However, many studies show that a single prediction model is insufficient to achieve very accurate predictions and affluent returns. In this paper, a novel portfolio construction approach is developed using a hybrid model based on machine learning for stock prediction and mean–variance (MV) model for portfolio selection. Specifically, two stages are involved in this model: stock prediction and portfolio selection. In the first stage, a hybrid model combining eXtreme Gradient Boosting (XGBoost) with an improved firefly algorithm (IFA) is proposed to predict stock prices for the next period. The IFA is developed to optimize the hyperparameters of the XGBoost. In the second stage, stocks with higher potential returns are selected, and the MV model is employed for portfolio selection. Using the Shanghai Stock Exchange as the study sample, the obtained results demonstrate that the proposed method is superior to traditional ways (without stock prediction) and benchmarks in terms of returns and risks.

As a type of uncertain differential equation, the uncertain vibration equation investigates the vibration of a spring-mass system whose external force is affected by an uncertain interference. The solution and inverse uncertainty distribution of the solution for a special type of uncertain spring vibration equation are derived, and an existence and uniqueness theorem of the solution is proved. Based on these results, the 'stability in measure' and 'stability in mean' are discussed for a general uncertain spring vibration equation.

This paper deals with a multiperiod multiobjective fuzzy portfolio selectiossn problem based on credibility theory. A credibilistic multiobjective mean-VaR model is formulated for the multiperiod portfolio selection problem, whereby the return is quantified by the credibilistic mean and the risk is measured by the credibilistic VaR. We also consider liquidity, cardinality, and upper and lower bound constraints to obtain a more realistic model. Furthermore, to solve the proposed model efficiently, an improved multiobjective bat algorithm termed IMBA is designed, in which three new strategies, i.e., the global best solution selection strategy, candidate solution generation strategy, and competitive learning strategy, are proposed to increase the convergence speed and improve the solution quality. Finally, comparative experiments are presented to show the applicability and superiority of the proposed approaches from two aspects. First, the designed IMBA is compared with seven typical algorithms, i.e., multiobjective particle swarm optimization, multiobjective artificial bee colony, multiobjective firefly algorithm, multiobjective differential evolution, multiobjective bat, the non-dominated sorting genetic algorithm (NSGA-II) and strength pareto evolutionary algorithm 2 (SPEA2), on a number of benchmark test problems. Second, the applicability of the proposed model to practical applications of portfolio selection is given under different circumstances.

Stock trend prediction is one of the most widely investigated and challenging problems for investors and researchers. Since the convolutional neural network (CNN) was introduced to analyze financial data, many researchers have dedicated to predicting stock trend by transforming stock market data into images. However, most of the existing studies just focused on individual stock information, and ignored stock market information, such as the existing correlations between stocks. In fact, the price volatility of a stock may be affected by those of other stocks, thus, taking the stock market information into the stock trend prediction can further improve the prediction performance. In this paper, we propose a novel method for stock trend prediction using graph convolutional feature based convolutional neural network (GC–CNN) model, in which both stock market information and individual stock information are considered. Specifically, an improved graph convolutional network (IGCN) and a Dual-CNN are designed to construct GC–CNN, which can simultaneously capture stock market features and individual stock features. Six randomly selected Chinese stocks are used to demonstrate the superior performance of the proposed GC–CNN based method. The experimental analysis demonstrates that the proposed GC–CNN based method outperforms several stock trend prediction methods and stock trading strategies.