Search Results (19)

Predicting financial stock returns remains challenging due to their inherent nonlinearity, non-stationarity, and sensitivity to market microstructure noise. Existing approaches typically rely on either econometric filtering techniques or deep learning models in isolation, limiting their ability to jointly capture latent dynamics and complex temporal dependencies. This study proposes a hybrid Extended Kalman Filter–Long Short-Term Memory (EKF–LSTM) framework that integrates nonlinear state-space filtering with deep sequential learning. The EKF component performs nonlinear state estimation and denoises to extract latent signals from noisy observations, while the LSTM network models nonlinear temporal dependencies in the filtered series. The proposed framework is evaluated using data from multiple international markets, including China, the United States, and Europe, providing cross-market evidence of model robustness. Empirical results show that the EKF–LSTM model consistently outperforms benchmark models (ARIMA, standalone EKF, LSTM, and GRU) across standard statistical metrics, including RMSE, MAE, and mean directional accuracy (MDA). In addition, the model delivers economically meaningful improvements under a long-only trading strategy, achieving higher risk-adjusted returns and lower maximum drawdowns relative to benchmark strategies. Diebold–Mariano tests further confirm that these performance gains are statistically significant. Overall, the findings demonstrate that integrating nonlinear state-space filtering with deep learning provides a robust and effective framework for financial time-series forecasting. However, the results should be interpreted with caution due to the limited sample size and the simplifying assumptions underlying the trading strategy. Full article

Stock trend forecasting remains challenging in real financial markets because data distributions evolve over time, and models trained under static settings often degrade during online deployment. Recent studies have introduced incremental and meta-incremental learning into stock forecasting, yet effective sequential adaptation remains constrained by two issues: financial multivariate time series require stronger representation modeling before downstream prediction, and repeated online updates may lead to forgetting and parameter drift. To address these issues, we propose ProAdapt, a bi-level meta-incremental learning framework for stock trend forecasting in non-stationary markets. ProAdapt contains two key components. The first is a Structural Spectral-Temporal Feature Adapter (SSTFA), which enhances financial time series representations by modeling non-uniform temporal importance and selective cross-factor interactions through adaptive soft window temporal encoding, frequency-domain structure modeling, and feature refinement. The second is online Elastic Weight Consolidation (EWC), which is incorporated into the outer-loop optimization to regularize sequential parameter updates and improve the balance between adaptation and stability. We evaluate ProAdapt on the CSI300 and CSI500 benchmarks under an incremental forecasting setting with sequential task updates. Experimental results across multiple backbones show that ProAdapt generally achieves favorable forecasting results relative to the compared baselines, with relatively clearer gains on CSI500. Additional ablation and analysis results further support the effectiveness of SSTFA and online EWC. Overall, the results suggest that combining explicit representation enhancement with stability-aware sequential updating is beneficial for incremental stock forecasting in evolving market environments. Full article

This study investigates regime-dependent forecasting of the Saudi stock market by combining macro-controlled dependence analysis with nonlinear predictive modeling. Using daily data from September 2010 to August 2025, we analyze the interaction between the Tadawul All Share Index (TASI) returns and crude oil returns while controlling for inflation and interest-rate dynamics. A four-variable VAR with macro controls is estimated separately in pre- and post-COVID regimes to characterize directional predictability and changes in transmission lags. We then evaluate out-of-sample return forecasting performance across econometric benchmarks (ARIMA, ARIMAX, and VAR) and machine learning models (LSTM and XGBoost) under a strictly time-ordered expanding-window design with sequential train/validation/test partitioning. The results indicate that traditional linear benchmarks exhibit limited predictive ability in both regimes, with negative out-of-sample explanatory power. By contrast, XGBoost delivers the strongest overall performance, achieving positive out-of-sample R² in both regimes (0.046 in pre-COVID and 0.010 in post-COVID), together with the lowest forecast errors (RMSE = 0.0081 pre-COVID; 0.0078 post-COVID). Interpretability analysis further reveals a regime-sensitive shift in drivers: short-horizon equity lag dynamics dominate during stable periods, whereas oil-related and macro-financial variables gain importance under turbulent conditions. Economic-value evaluation supports the practical relevance of these gains, showing that XGBoost-based signals yield superior risk-adjusted trading outcomes and remain favorable under downside-risk and drawdown-based assessment. Overall, these findings highlight that forecasting in oil-linked emerging markets is inherently regime-dependent and that nonlinear ensemble learners, particularly XGBoost, provide a more robust and economically meaningful approach under structural change. Full article

Stock return prediction for quantitative trading in U.S. equity markets has evolved from parametric econometric modeling toward data-driven deep learning systems that must jointly capture temporal dynamics, discontinuous jumps, and evolving cross-asset dependencies. Existing approaches still face three key challenges in deep learning-based stock return prediction: jump-aware temporal modeling is often missing or handled by ad hoc heuristics; higher-order stock relations are frequently encoded by static graphs/hypergraphs that do not adapt across market conditions, and temporal and relational learning are commonly implemented as sequential blocks with limited bidirectional interaction. We propose LévyHyper, an end-to-end framework that unifies jump-aware temporal encoding with regime-adaptive dynamic hypergraph learning and multi-scale hypergraph reasoning. LévyHyper integrates a neural jump-aware temporal layer motivated by Lévy jump-diffusion modeling, a regime-weighted fusion of predefined and learned hyperedges via a differentiable constructor, and a multi-scale hypergraph convolution module for hierarchical temporal aggregation. Experiments on S&P 500 data (463 stocks, 10 evaluation phases, prediction horizon $\tau =5$ trading days) show that LévyHyper improves IC/RankIC and portfolio-level Sharpe ratio over strong baselines on average. We additionally report uncertainty estimates, significance tests, and transaction-cost sensitivity to support robust conclusions. Full article

Deep learning has transformed numerous areas of data science by achieving outstanding performance in tasks such as image recognition, speech processing, and natural language understanding. Recently, the challenges of financial forecasting—marked by nonlinear dynamics, volatility, and regime shifts—have attracted increasing attention from the deep learning community. Among these approaches, Convolutional Neural Networks (CNNs), originally developed for spatial data, have shown strong potential for modelling financial time series. This study presents an interpretable CNN-based framework for stock price forecasting using the S&P 500 index as a case study. The proposed approach integrates historical price data with technical indicators within a unified experimental design and compares performance against traditional statistical (ARIMA) and sequential deep learning (LSTM) baselines. Empirical results demonstrate that the CNN model achieves superior predictive Accuracy while maintaining computational efficiency and interpretability through SHAP and Grad-CAM analyses. The findings suggest that lightweight CNN architectures can serve as effective, transparent tools for short-horizon financial forecasting, and future research may extend this framework to multimodal settings incorporating sentiment or news-based data. Full article

With agriculture being the second biggest contributor to greenhouse gas (GHG) emissions through the excessive use of fertilizers, machinery, and inefficient farming practices, global efforts to reduce emissions have been intensified, opting for smarter, data-driven solutions. However, while machine learning (ML) offers powerful predictive capabilities, its black-box nature presents a challenge for trust and adoption, particularly when integrated with auditable financial technology (FinTech) principles. To address this gap, this work introduces a novel, explanation-focused GHG emission optimization framework for IoT-enabled smart agriculture that is both transparent and prescriptive, distinguishing itself from macro-level land-use solutions by focusing on optimizable management practices while aligning with core FinTech principles and pollutant stock market mechanisms. The framework employs a two-stage statistical methodology that first identifies distinct agricultural emission profiles from macro-level data, and then models these emissions by developing a cluster-oriented principal component regression (PCR) model, which outperforms simpler variants by approximately 35% on average across all clusters. This interpretable model then serves as the core of a FinTech-aligned optimization framework that combines cluster-oriented modeling knowledge with a sequential least squares quadratic programming (SLSQP) algorithm to minimize emission-related costs under a carbon pricing mechanism, showcasing forecasted cost reductions as high as 43.55%. Full article

Predicting stock prices is a popular area of study within the realms of data mining and machine learning. Precise forecasting can assist investors in mitigating the risks associated with their investments. Given the unpredictable nature of the stock market, influenced by policy changes, stock data often display high levels of fluctuation and randomness, aligning closely with the prevailing market sentiment. Moreover, diverse datasets related to stocks are rich in historical data that can be leveraged to forecast future trends. However, traditional forecasting models struggle to harness this information effectively, which restricts their predictive capabilities and accuracy. To improve the existing issues, this research introduces a novel stock prediction model based on a deep-learning neural network, named after EL-MTSA, which leverages the multifaceted characteristics of stock data along with ensemble learning optimization. In addition, a new evaluation index via market-wide sentiment analysis is designed to enhance the forecasting performance of the stock prediction model by adeptly identifying the latent relationship between the target stock index and dynamic market sentiment factors. Subsequently, many demonstration experiments were conducted on three practical stock datasets, the CSI 300, SSE 50, and CSI A50 indices, respectively. Experiential results show that the proposed EL-MTSA model has achieved a superior predictive performance, surpassing various comparison models. In addition, the EL-MTSA can analyze the impact of market sentiment and media reports on the stock market, which is more consistent with the real trading situation in the stock market, and indicates good predictive robustness and credibility. Full article

Stock market forecasting is a critical area in financial research, yet the inherent volatility and non-linearity of financial markets pose significant challenges for traditional predictive models. This study proposes a hybrid deep learning model, integrating Long Short-Term Memory (LSTM) networks and Convolutional Neural Networks (CNNs) with technical indicators to enhance the predictive accuracy of stock price movements. The model is evaluated using Root Mean Squared Error (RMSE), Mean Absolute Error (MAE), Mean Absolute Percentage Error (MAPE), and R² score on the S&P 500 index over a 14-year period. Results indicate that the LSTM-CNN hybrid model achieves superior predictive performance compared to traditional models, including Support Vector Machines (SVMs), Random Forest (RF), and ARIMAs, by effectively capturing both long-term trends and short-term fluctuations. While Random Forest demonstrated the highest raw accuracy with the lowest RMSE (0.0859) and highest R² (0.5655), it lacked sequential learning capabilities. The LSTM-CNN model, with an RMSE of 0.1012, MAE of 0.0800, MAPE of 10.22%, and R² score of 0.4199, proved to be highly competitive and robust in financial time series forecasting. The study highlights the effectiveness of hybrid deep learning architectures in financial forecasting and suggests further enhancements through macroeconomic indicators, sentiment analysis, and reinforcement learning for dynamic market adaptation. It also improves risk-aware decision-making frameworks in volatile financial markets. Full article

The goal of this research is to introduce and thoroughly investigate a new methodology for the assessment of sequential regularity in volatility time series. Three volatility estimators based on daily range data are analyzed: (1) the Parkinson estimator, (2) the Garman–Klass estimator, and (3) the Rogers–Satchell estimator. To measure the level of complexity of time series, the modified Shannon entropy based on symbol-sequence histograms is utilized. Discretization of the time series of volatility changes into a sequence of symbols is performed using a novel encoding procedure with two thresholds. Five main stock market indexes are analyzed. The whole sample covers the period from January 2017 to December 2023 (seven years). To check the robustness of our empirical findings, two sub-samples of equal length are investigated: (1) the pre-COVID-19 period from January 2017 to February 2020 and (2) the COVID-19 pandemic period from March 2020 to April 2023. An additional formal statistical analysis of the symbol-sequence histograms is conducted. The empirical results for all volatility estimators and stock market indexes are homogeneous and confirm that the level of regularity (in terms of sequential patterns) in the time series of daily volatility changes is high, independently of the choice of sample period. These results are important for academics and practitioners since the existence of regularity in the time series of volatility changes implies the possibility of volatility prediction. Full article

Stock rank prediction is an important and challenging task. Recently, graph-based prediction methods have emerged as a valuable approach for capturing the complex relationships between stocks. Existing works mainly construct static undirected relational graphs, leading to two main drawbacks: (1) overlooking the bidirectional asymmetric effects of stock data, i.e., financial messages affect each other differently when they occur at different nodes of the graph; and (2) failing to capture the dynamic relationships of stocks over time. In this paper, we propose a Multi-modal Temporal Dynamic Graph method (MTDGraph). MTDGraph comprehensively considers the bidirectional relationships from multi-modal stock data (price and texts) and models the time-varying relationships. In particular, we generate the textual relationship strength from the topic sensitivity and the text topic embeddings. Then, we inject a causality factor via the transfer entropy between the interrelated stock historical sequential embeddings as the historical relationship strength. Afterwards, we apply both the textual and historical relationship strengths to guide the multi-modal information propagation in the graph. The framework of the MTDGraph method consists of the stock-level sequential embedding layer, the inter-stock relation embedding layer based on temporal dynamic graph construction and the multi-model information fusion layer. Finally, the MTDGraph optimizes the point-wise regression loss and the ranking-aware loss to obtain the appropriate stock rank list. We empirically validate MTDGraph in the publicly available dataset, CMUN-US and compare it with state-of-the-art baselines. The proposed MTDGraph method outperforms the baseline methods in both accuracy and investment revenues. Full article

Generating high-quality synthetic data is essential for advancing machine learning applications in financial time series, where data scarcity and privacy concerns often pose significant challenges. This study proposes a novel hybrid architecture that combines variational autoencoders (VAEs) with Markov Chain Monte Carlo (MCMC) sampling to enhance the generation of robust synthetic sequential data. The model leverages Gated Recurrent Unit (GRU) layers for capturing long-term temporal dependencies and MCMC sampling for effective latent space exploration, ensuring high variability and accuracy. Experimental evaluations on datasets of Google, Tesla, and Nestlé stock prices demonstrate the model’s superior performance in preserving statistical and temporal patterns, as validated by quantitative metrics (discriminative and predictive scores), statistical tests (Kolmogorov–Smirnov), and t-Distributed Stochastic Neighbour Embedding (t-SNE) visualisations. The experiments reveal the model’s scalability, maintaining high fidelity even under augmented dataset sizes and missing data scenarios. These findings position the proposed framework as a computationally efficient and structurally simple alternative to Generative Adversarial Network (GAN)-based methods, suitable for real-world applications in data-driven financial modelling. Full article

Forecasting stock markets is challenging due to the influence of various internal and external factors compounded by the effects of globalization. This study introduces a data-driven approach to forecast S&P 500 returns by incorporating macroeconomic indicators including gold and oil prices, the volatility index, economic policy uncertainty, the financial stress index, geopolitical risk, and shadow short rate, with ten technical indicators. We propose three hybrid deep learning models that sequentially combine convolutional and recurrent neural networks for improved feature extraction and predictive accuracy. These models include the deep belief network with gated recurrent units, the LeNet architecture with gated recurrent units, and the LeNet architecture combined with highway networks. The results demonstrate that the proposed hybrid models achieve higher forecasting accuracy than the single deep learning models. This outcome is attributed to the complementary strengths of convolutional networks in feature extraction and recurrent networks in pattern recognition. Additionally, an analysis using the Shapley method identifies the volatility index, financial stress index, and economic policy uncertainty as the most significant predictors, underscoring the effectiveness of our data-driven approach. These findings highlight the substantial impact of contemporary uncertainty factors on stock markets, emphasizing their importance in studies analyzing market behaviour. Full article

Fungal genetic systems ideally combine molecular tools for genome manipulation and a sexual reproduction system to create an informative assortment of combinations of genomic modifications. When employing the sexual cycle to generate multi-mutants, the background genotype variations in the parents may result in progeny phenotypic variation obscuring the effects of combined mutations. Here, to mitigate this variation in Fusarium verticillioides, we generated a MAT1-2 strain that was near isogenic to the sequenced wild-type MAT1-1 strain, FGSC7600. This was accomplished by crossing FGSC7600 with the divergent wild-type MAT1-2 strain FGSC7603 followed by six sequential backcrosses (e.g., six generations) of MAT1-2 progeny to FGSC7600. We sequenced each generation and mapped recombination events. The parental cross involved twenty-six crossovers on nine of the eleven chromosomes. The dispensable chromosome 12, found in FGSC7603 but lacking in FGSC7600, was not present in the progeny post generation five. Inheritance of complete chromosomes without crossover was frequently observed. A deletion of approximately 140 kilobases, containing 54 predicted genes on chromosome 4, occurred in generation 4 and was retained in generation 5 indicating that these genes are dispensable for growth and both asexual and sexual reproduction. The final MAT1-2 strain TMRU10/35 is about 93% identical to FGSC7600. TMRU10/35 is available from the Fungal Genetics Stock Center as FGSC27326 and from the ARS Culture Collection as NRRL64809. Full article

The stock market, characterised by its complexity and dynamic nature, presents significant challenges for predictive analytics. This research compares the effectiveness of neural network models in predicting the S&P500 index, recognising that a critical component of financial decision making is market volatility. The research examines neural network models such as Long Short-Term Memory (LSTM), Convolutional Neural Network (CNN), Artificial Neural Network (ANN), Recurrent Neural Network (RNN), and Gated Recurrent Unit (GRU), taking into account their individual characteristics of pattern recognition, sequential data processing, and handling of nonlinear relationships. These models are analysed using key performance indicators such as the Root Mean Square Error (RMSE), Mean Absolute Percentage Error (MAPE), and Directional Accuracy, a metric considered essential for prediction in both the training and testing phases of this research. The results show that although each model has its own advantages, the GRU and CNN models perform particularly well according to these metrics. GRU has the lowest error metrics, indicating its robustness in accurate prediction, while CNN has the highest directional accuracy in testing, indicating its efficiency in data processing. This study highlights the potential of combining metrics for neural network models for consideration when making decisions due to the changing dynamics of the stock market. Full article

The stock price is a culmination of numerous factors that are not necessarily quantifiable and significantly affected by anomalies. The forecasting accuracy of stock prices is negatively affected by these anomalies. However, very few methods are available for detecting, modelling, and compensating for anomalies in financial time series given the critical roles of better management of funds and accurate forecasting of anomalies. Time series data are a data type that changes over a defined time interval. Each value in the data set has some relation to the previous values in the series. This attribute of time series data allows us to predict the values that will follow in the series. Typical prediction models are limited to following the patterns in the data set without being able to compensate for anomalous periods. This research will attempt to find a machine learning method to detect outliers and then compensate for these detections in the prediction made. This concept was previously unimplemented, and therefore, it will make use of theoretical work on market forecasting, outliers and their effects, and machine learning methods. The ideas implemented in the paper are based upon the efficient market hypothesis (EMH), in which the stock price reflects knowledge about the market. The EMH hypothesis cannot account for consumer sentiment towards a stock. This sentiment could produce anomalies in stock data that have a significant influence on the movement of the stock market. Therefore, the detection and compensation of outliers may improve the predictions made on stock movements. This paper proposes a deep learning method that consists of two sequential stages. The first stage is an outlier detection model based on a long short-term memory (LSTM) network auto-encoder that can determine if an outlier event has occurred and then create an associated value of this occurrence for the next stage. The second stage of the proposed method uses a higher-order neural network (HONN) model to make a prediction based on the output of the first stage and the stock time series data. Real stock data and standalone prediction models are used to validate this method. This method is superior at predicting stock time series data by compensating for outlier events. The improvement is quantifiable if the data set contains an adequate amount of anomalous periods. We may further apply the proposed method of compensating for outliers in combination with other financial time series prediction methods to offer further improvements and stability. Full article