Prediction of Stock Prices using LSTM Model and Comparison with ML Algorithms

Abstract

Guessing stock rates for the future is one of the key components of financial deals market analysis, since they can provide information that acts like risk management and investing alternatives. Deep learning’s (DLs) strong processing capabilities and aptitude to understand the non-technical, non- linear interactions have been able to establish considerable progress in stock price forecast. The LSTM networks are good for the forecasting in the problems related to time series such as stock price forecast since they have the capability of taking into account lagged effects and avoiding prospects like gradients that vanish. After considering the above performance matrix, which are the following: accuracy, precision, recall, and the F1-score, it will be possible for the models to be evaluated completely. These criteria give a comprehensive evaluation of the predictive accuracy, the robustness, and the generalization ability of the algorithms. Mind the results to show you that traditional ML methods, which can provide prompt and simple interpretable results, the LSTM-based model makes a better job at modeling the complex temporal patterns found in stock price data. The relative research sets out how DL can raise stock price prediction accuracy thus it enriches the financial analysts and investors with new insights. Our suggested LSTM model delivers 15.78% more accurate results than Linear Regression and 6.91% more accurate results than Random Forest, thus, it convinces everyone about its highlevel performance in stock price prediction.

Introduction

The stock price represents the current market value at which a company’s share is traded and constantly changes due to market supply and demand. It increases when demand exceeds supply and decreases when supply exceeds demand. Several factors influence stock prices, including company financial performance, economic conditions, investor sentiment, and overall market trends. Because of this volatility, predicting stock prices is essential for investment decisions and risk management.

Traditional statistical and machine learning (ML) models have been used for stock prediction, but Deep Learning (DL), particularly Long Short-Term Memory (LSTM) networks, has shown superior performance. LSTMs are a type of Recurrent Neural Network (RNN) designed to handle sequential and time-series data while overcoming issues like vanishing gradients. Their ability to retain long-term dependencies and capture complex, nonlinear patterns makes them especially suitable for financial forecasting.

Research shows that LSTM models consistently outperform traditional ML models such as Linear Regression (LR), Random Forest (RF), Decision Trees (DT), and Support Vector Machines (SVM) in stock price prediction tasks. LSTMs are particularly effective at recognizing temporal patterns and adapting to changing market conditions. However, they require large amounts of high-quality data and are prone to overfitting if not properly regularized.

The proposed methodology involves collecting historical stock data (e.g., from Yahoo Finance), preprocessing it using normalization and sliding window techniques, and training an LSTM model optimized with Adam. The model’s performance was compared with Linear Regression and Random Forest using TCS stock data.

Results show that LSTM significantly outperformed traditional models, achieving:

• Accuracy: 89.34%

• Precision: 90.21%

• Recall: 88.76%

• F1 Score: 89.48%

Compared to Random Forest and Linear Regression, LSTM demonstrated better temporal dependency handling, improved precision-recall balance, and stronger overall predictive reliability. These findings highlight the importance of deep learning—particularly LSTM networks—in enhancing stock price forecasting, improving risk management, and supporting informed investment decisions.

Conclusion

The work that were accomplished demonstrated that these models have the transformative potential in the prediction of stock prices, and thus made a significant leap for financial forecasting technologies. The LSTM model manages to achieve an impressive 89.34% success rate and to be more efficient than other basal ML models like LR and RF model, which are important reason of existing and making a claim of the LSTM model’s superiority in the ability to capture complex temporal dependencies and market dynamics. The investigators comprehensive analysis put radar on the models unmatched ability to deal with complex stock market trends and to represent values such as 90.21 precision and an F1 Score of 89.48%, thus pointing out the effectiveness of the model in forecasting capabilities. The findings present as a contribution in the academic understanding of DL applications in financial forecasting and as well as are understandable for investors, and financial analysts in terms of getting more sophisticated prediction techniques. Moreover, the research indicates that, in the future, the proposed promising directions such as data improvement, the design of a hybrid model with a mix of traditional and advanced methods, and DL model tuning and refine. will be the judgments of further exploration. In the end, this study, first of all, proofs the possibilities that advanced ML techniques might bring in helping our understanding and prediction of the stock market behavior, although it also reminders that a cautious and multifaceted approach investment strategies are necessary to be present.

References

[1] Stock Price - Definition, Price Changes, How to Determine [2] What is a Share Price? | Definition and Explanation | IG International [3] Factors That Move Stock Prices Up and Down [4] Stock Market Price Prediction Using Machine Learning [5] Rath, S., Das, N.R., & Pattanayak, B.K. (2023). An Analytic Review on Stock Market Price Prediction using Machine Learning and Deep Learning Techniques. Recent Patents on Engineering. [6] Assistant, P.P., Shravani, S., Tajuddin, M., & Srija, R. (2023). A Model to Stock Price Prediction using Deep Learning. 2023 International Conference on Sustainable Computing and Smart Systems (ICSCSS), 242-252. [7] Gautam, B., Kandel, S., Shrestha, M. and Thakur, S. (2024) Comparative Analysis of Machine Learning Models for Stock Price Prediction: Leveraging LSTM for Real-Time Forecasting. Journal of Computer and Communications, 12, 52-80. [8] Perner, P. (Ed.). (2007). Machine Learning and Data Mining in Pattern Recognition: 5th International Conference, MLDM 2007, Leipzig, Germany, July 18-20, 2007, Proceedings (Vol. 4571). Springer Science & Business Media. [9] https://towardsdatascience.com/predicting-stock-prices-using-a- keras- lstm-model-4225457f0233 [10] JIANG, W. (2021). Applications of deep learning in stock market prediction: recent progress. Expert Systems with Applications, 184, 115537. [11] Gulmez, B. (2023). Stock price prediction with optimized deep LSTM network with artificial rabbits optimization algorithm. Expert Systems with Applications, 227, 120346. [12] Hochreiter, S. (1997). Long Short-term Memory. Neural Computation MIT-Press. [13] Sen, J. (2022). Analysis and Forecasting of Financial Time Series: Selected Cases. Cambridge Scholars Publishing. [14] Afzal, F., Haiying, P., Afzal, F., & Nazir, M. Emergence of Artificial Intelligence and Machine Learning Tools in Predicting Stock Market Performance: A Scientometric View. Available at SSRN 4637914. [15] Zhang, G., & Huang, J. (2021). Exploring the Efficacy of Machine Learning in Stock Price Prediction: A Comparative Study. Expert Systems with Applications, 165, 113926. [16] Ali, F., & Suri, P. (2022). A Bibliometric Analysis of Artificial Intelligence-Based Stock Market Prediction. The Eurasia Proceedings of Educational and Social Sciences, 27, 17-35. [17] Han, Z., Zhu, X., & Su, Z. (2024). Forecasting Maritime and Financial Market Trends: Leveraging CNN- LSTM Models for Sustainable Shipping and China’s Financial Market Integration. Sustainability, 16(22), 9853 [18] Kumar, N., & Singh, A. (2020). Assessment of Machine Learning and Deep Learning Models for Stock Market Forecasting. Decision Support Systems, 141, 113464. [19] Zhao, R., Chen, Z., & O’Reilly, M. (2022). A Comprehensible Review of Time Series Prediction Techniques with Reinforcement Learning and LSTM. Machine Learning and Data Mining in Pattern Recognition, 127-140. [20] Zhao, W., Lu, J., & Wang, G. (2022). A Comprehensive Review of Deep Learning in Stock Market Prediction. Expert Systems with Applications, 183, 115368. [21] Zhang, G., Eddy Patuwo, B., & Hu, M. Y. (1998). Forecasting with Artificial Neural Networks: The State of the Art. International Journal of Forecasting, 14(1), 35–62. [22] Bollen, J., Mao, H., & Zeng, X. (2011). Twitter Mood Predicts the Stock Market. Journal of Computational Science, 2(1), 1–8.

Copyright

Copyright © 2026 Utkarsh Pandey, Abhishek , Meenakshi Sharma. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.