Nasya Vhazira, Irmayatul Hikmah, Mas Aly Afandi

Performance Analysis of Quantum Long Short-Term Memory (QLSTM) Models for TLKM Stock Price Prediction

Introduction

Performance analysis of quantum long short-term memory (qlstm) models for tlkm stock price prediction. Evaluate Quantum Long Short-Term Memory (QLSTM) models for TLKM stock price prediction. Systematically analyzes optimizers, epochs, and hidden units, finding Adam provides stable, accurate forecasts.

Abstract

Stock price prediction is a challenging task due to its nonlinear, dynamic, and temporal characteristics, yet accurate forecasting models are crucial for decision-making in volatile stocks such as PT Telkom Indonesia Tbk (TLKM). Despite the rapid adoption of AI-based forecasting methods, several research gaps remain. Empirical studies on Quantum Long Short-Term Memory (QLSTM) are still relatively limited compared to classical LSTM variants, particularly for emerging market datasets. Existing research also tends to emphasize architectural comparisons rather than systematically analyzing training configurations. The joint effects of optimizer selection, epoch number, and hidden unit size on QLSTM performance have not been comprehensively evaluated, and many studies rely on limited evaluation metrics, reducing the strength of robustness assessment. To address these gaps, this study applies a QLSTM model to predict stock opening prices using historical time-series data and systematically evaluates the impact of different optimizers. The model is trained using Adam, Nadam, RMSprop, and SGD with epoch variations (50–250) and hidden units (8, 16, 32). Performance is measured using accuracy, MAE, MSE, RMSE, MAPE, and R² to ensure a comprehensive evaluation. The results indicate that adaptive optimizers consistently outperform SGD, with Adam providing the most stable and accurate predictions, highlighting the importance of optimizer choice and hyperparameter configuration in QLSTM-based stock forecasting.

Review

This paper tackles the intricate problem of stock price prediction, specifically focusing on the volatile PT Telkom Indonesia Tbk (TLKM) stock. It effectively identifies several pertinent research gaps in the application of Quantum Long Short-Term Memory (QLSTM) models, particularly the limited empirical studies in emerging markets, the lack of systematic analysis into training configurations, and insufficient use of comprehensive evaluation metrics. The study's core strength lies in its systematic evaluation of optimizer selection, epoch number, and hidden unit size on QLSTM performance, utilizing a robust suite of six distinct metrics (Accuracy, MAE, MSE, RMSE, MAPE, R²). This methodical approach directly addresses the identified shortcomings in existing literature, providing a strong empirical foundation for optimizing QLSTM models for financial time-series forecasting. While the study makes a commendable effort in systematically evaluating QLSTM training configurations, a potential area for enhancement not detailed in the abstract concerns the direct comparison against classical LSTM or other state-of-the-art classical deep learning models. Although the paper acknowledges the limited empirical studies on QLSTM versus classical variants as a research gap, the abstract primarily focuses on hyperparameter tuning within QLSTM itself. Without a direct performance benchmark against established classical models, it becomes challenging to fully ascertain the unique advantages or specific performance uplift contributed by the quantum components of the QLSTM architecture for this particular forecasting task. Further, while "historical time-series data" is mentioned, specifics regarding its temporal range, frequency, and any unique preprocessing steps applied would enhance reproducibility and contextual understanding. In conclusion, this paper offers a significant and practical contribution to the emerging domain of quantum machine learning in finance. Its systematic analysis of QLSTM hyperparameters and the clear demonstration of adaptive optimizers, particularly Adam, as critical for stable and accurate predictions, provides valuable guidance for both researchers and practitioners. The findings effectively highlight the importance of meticulous hyperparameter tuning in maximizing the potential of quantum-enhanced models. Future research could build upon this foundation by explicitly quantifying the performance gains of QLSTM against classical benchmarks, exploring the impact of various quantum feature encoding strategies, and extending the analysis to a wider array of volatile stocks or diverse market conditions to further solidify the model's generalizability and practical utility.

Full Text

Comments

You need to be logged in to post a comment.