Financial time series forecasting remains one of the most challenging tasks due to its noisy, non-stationary, and stochastic nature. The paper represents the empirical evaluation of a multi-input Convolutional Neural Network regarding stock market trend classification. The novelty of our approach is based on feature engineering: transforming one-dimensional technical indicators (namely, 20-day simple moving average and Bollinger Bands) into two-dimensional image representations by means of Quantum Gramian Angular Field encoding technique. This is developed by leveraging a simulated quantum circuit mapping time series into the feature space designed to preserve the temporal dynamics. The produced QGAF images are fed into CNN in parallel, where the model is trained to classify whether the stock will go up or down the following day. The performance evaluation is conducted on a historical dataset of US major equities. On one hand, empirical results manifest enormous overfitting and provide modest test accuracy of$\mathbf{5 4. 2 \%}$. At the same time, a deep analysis provided including precision and recall metrics testifies about strong bias of the model. In this way, the study confirms the QGAF encoding pipeline as suitable for feature engineering technique and creates an important empirical baseline for the future. The main contribution of the paper is given by the critical, in-depth evaluation of this hybrid quantum-classical, multi-input architecture, outlining the enormous challenges still to be overcome when applying such techniques to financial data.
Financial time series forecasting remains one of the most challenging tasks due to its noisy, non-stationary, and stochastic nature. The paper represents the empirical evaluation of a multi-input Convolutional Neural Network regarding stock market trend classification. The novelty of our approach is based on feature engineering: transforming one-dimensional technical indicators (namely, 20-day simple moving average and Bollinger Bands) into two-dimensional image representations by means of Quantum Gramian Angular Field encoding technique. This is developed by leveraging a simulated quantum circuit mapping time series into the feature space designed to preserve the temporal dynamics. The produced QGAF images are fed into CNN in parallel, where the model is trained to classify whether the stock will go up or down the following day. The performance evaluation is conducted on a historical dataset of US major equities. On one hand, empirical results manifest enormous overfitting and provide modest test accuracy of54.2%. At the same time, a deep analysis provided including precision and recall metrics testifies about strong bias of the model. In this way, the study confirms the QGAF encoding pipeline as suitable for feature engineering technique and creates an important empirical baseline for the future. The main contribution of the paper is given by the critical, in-depth evaluation of this hybrid quantum-classical, multi-input architecture, outlining the enormous challenges still to be overcome when applying such techniques to financial data.
Quantum machine learning offers a promising pathway for enhancing stock market prediction, particularly under complex, noisy, and highly dynamic financial environments. However, many classical forecasting models struggle with noisy input, regime shifts, and limited generalization capacity. To address these challenges, we propose a Quantum Temporal Convolutional Neural Network (QTCNN) that combines a classical temporal encoder with parameter-efficient quantum convolution circuits for cross-sectional equity return prediction. The temporal encoder extracts multi-scale patterns from sequential technical indicators, while the quantum processing leverages superposition and entanglement to enhance feature representation and suppress overfitting. We conduct a comprehensive benchmarking study on the JPX Tokyo Stock Exchange dataset and evaluate predictions through long-short portfolio construction using out-of-sample Sharpe ratio as the primary performance metric. QTCNN achieves a Sharpe ratio of 0.538, outperforming the best classical baseline by approximately 72\%. These results highlight the practical potential of quantum-enhanced forecasting model, QTCNN, for robust decision-making in quantitative finance.
Stock market forecasting remains one of the most computationally demanding tasks in financial engineering, requiring sophisticated models capable of processing nonlinear patterns, high-frequency volatility, and large volumes of time-series data. Recent advancements in quantum computing offer the potential to surpass classical machine learning (ML) by leveraging principles such as superposition and entanglement to accelerate learning and optimization. This paper conducts a comprehensive comparison between quantum machine learning (QML) models and classical deep learning architectures for stock price prediction, emphasizing the technological improvements and optimization of computational methods. Our analysis integrates insights from prior studies, including the application of quantum-enhanced Gramian Angular Fields with CNNs [2], shallow parameterized quantum circuits for market indicators [4], and hybrid quantum-classical algorithms tailored to finance [1]. Experimental evaluation (using IBM Qiskit's quantum simulators and real hardware) demonstrates competitive performance of quantum neural networks (QNNs) against state-of-the-art recurrent neural networks (RNNs) and long short-term memory (LSTM) architectures [3]. Across AAPL/MSFT and short horizons, compact QNNs match strong LSTM baselines while using an order of magnitude fewer parameters, and—critically—transfer from calibrated simulators to IBM hardware with marginal degradation after mitigation. This device-aware training/evaluation loop exemplifies a deployable path for financial QML under realistic noise and resource constraints and motivates standardized, time-series-aware testing for the community. Results highlight the engineering value of quantum acceleration in feature encoding, the research novelty of hybrid ensemble strategies, and the future scope of benchmarking across leading quantum computing machines such as IBM, Google Sycamore, Rigetti, IonQ, and Xanadu.
This research explores a novel hybrid neural network architecture for stock price forecasting that integrates Variational Quantum Layers (VQLs) to enhance efficiency and trainability. It examines various quantum-classical algorithms through extensive simulations on both classical and quantum hardware, utilizing key stock indicators like MACD, EMA, Bollinger Bands, and momentum metrics, alongside sentiment analysis of financial news. The dataset used includes real-time stock prices for 15 companies, focusing on Merck & Co. (MRK). By framing stock price forecasting as a binary classification problem, It proposes QLSTM4FM i.e. a unique framework based on quantum-assisted LSTM model, which significantly enhances stock price forecasting. It conducts a thorough and comprehensive performance evaluation of the proposed framework against classical LSTM models. Results show that QLSTM4FM, featuring quantum-enhanced layers, outperforms classical LSTM models in accuracy, lower loss, and faster convergence, demonstrating the potential of quantum computing in financial forecasting.
The need for smart, scalable algorithms to predict stock market movements has grown in response to the deluge of financial data coming from disparate sources including news articles, social media, and market pricing. Despite its success, traditional deep learning is ill-equipped to handle the complex nonlinear relationships and high-dimensional interactions found in financial markets. To get around these problems, this study presents a paradigm for quantum-enhanced multimodal intelligence that uses deep learning and quantum computing to predict stock market trends. Through the use of quantum-inspired feature mapping and quantumassisted optimisation, modality-specific encoders create representations from numerical market indicators and textual sentiment data. Both representation learning and exploring solution spaces are improved by this hybrid design. When tested on industry-standard financial datasets, the suggested model achieved better accuracy, precision, recall, and F1-score than typical multimodal deep learning baselines. The results of ablation studies further demonstrate that combining multimodal fusion with quantum enhancement improves the accuracy of predictions. Based on the findings, next-gen financial forecasting systems may incorporate quantum-augmented intelligence.
Prediction of stock price remains a challenging task because financial markets are dynamic and often non-linear, unpredictable patterns. To deal with such unpredictable data, the architecture that’s created must identify both time based trends and hidden patterns in data. In this study, we explore a hybrid modelling approach by combining both the classical deep learning techniques with quantum computing principles. Specifically, a Long Short-Term Memory (LSTM) is combined with a Quantum Neural Network (QNN) to understand the benefits of using quantum in time series forecasting. Rather than depending on a single quantum configuration, multiple feature maps and ansatz circuit designs were evaluated to identify the most suitable arrangement for the QLSTM model. Resultant framework blends ideas from both classical computing and quantum approaches to improve how we predict patterns over time in real-life scenarios.
The rapid evolution of financial technology has driven the adoption of advanced computational techniques for stock market prediction. Traditional statistical models often fail to capture the nonlinear, stochastic, and high-dimensional nature of financial time series, leading to increased interest in hybrid approaches combining deep learning and quantum computing. This review presents an optimized equivariant split attention quantum neural network framework for stock market prediction and recommendation systems. The proposed approach integrates quantum neural networks with equivariant architectures to enhance feature representation and generalization, while split attention mechanisms improve temporal dependency modeling and computational efficiency. The study explores optimization techniques such as quantum circuit pruning, adaptive hyperparameter tuning, and hardware-aware neural architecture search. It evaluates performance using benchmark datasets from major stock exchanges, incorporating price data, macroeconomic indicators, and sentiment features. Results indicate improved prediction accuracy, reduced error, and enhanced risk-adjusted returns compared to conventional models such as LSTM and transformer architectures. Overall, the framework demonstrates strong potential for real-time trading applications, offering a scalable and efficient solution for next-generation intelligent financial forecasting systems.