INTERNATIONAL JOURNAL OF LATEST TECHNOLOGY IN ENGINEERING,  
MANAGEMENT & APPLIED SCIENCE (IJLTEMAS)  
ISSN 2278-2540 | DOI: 10.51583/IJLTEMAS | Volume XIV, Issue XII, December 2025  
Hybrid AI Models for Real-Time Stock Management and Market  
Price Prediction  
1 Sumit Kumar, 1 Manoj Kumar, 1 Sharad Kumar, 1 Sachin Kumar, 1 Jagdeep Singh, 2 Vikas Sharma  
1 School of Engineering & Technology, Shri Venkateshwara University, Gajraula, U.P. India  
2 Department of Computer Applications, SRM Institute of Science and Technology, Delhi NCR Campus,  
Ghaziabad, U.P. India  
DOI : https://doi.org/10.51583/IJLTEMAS.2025.1412000108  
Received: 19 December 2025; Accepted: 24 December 2025; Published: 10 January 2026  
ABSTRACT  
Efficient stock management and accurate market price prediction are essential in dynamic and volatile business  
environments, where traditional forecasting and inventory control methods often lack adaptability and real-  
time responsiveness. This paper proposes a hybrid artificial intelligence (AI) framework for real-time stock  
management and market price prediction, integrating machine learning and deep learning models to capture  
both linear trends and nonlinear market patterns. The system processes real-time transactional data, historical  
stock prices, and relevant market indicators to continuously update inventory levels and forecast future price  
movements. Feature engineering, data normalization, and model fusion techniques are employed to enhance  
prediction accuracy and robustness. A decision-support module utilizes predicted demand and price trends to  
optimize inventory replenishment, reduce stockouts, and minimize overstocking costs. Experimental  
evaluation using real-world market datasets demonstrates that the proposed hybrid model outperforms  
individual predictive approaches in terms of forecasting accuracy, adaptability, and inventory efficiency, as  
reflected by improved MAE and RMSE values. The results confirm the effectiveness of the proposed approach  
as a scalable and intelligent solution for real-time stock management and market price prediction applications.  
Experimental results show that the proposed RT-HAF model reduces RMSE by approximately 23% compared  
to LSTM and improves inventory service level by over 10%.  
Keywords—Hybrid AI, Real-Time Stock Management, Market Price Prediction, Machine Learning, Deep  
Learning, Time-Series Forecasting, Predictive Analytics, Inventory Optimization, Decision Support Systems.  
INTRODUCTION  
The rapid growth of digital commerce, algorithmic trading, and data-driven decision-making has significantly  
increased the complexity of stock management and market price forecasting in modern business environments.  
Organizations operating in retail, supply chain management, and financial markets are required to manage  
large volumes of inventory while simultaneously responding to frequent price fluctuations driven by demand  
variability, economic indicators, and market sentiment. Traditional stock management systems, which rely on  
static thresholds and historical averages, are often inadequate for handling real-time data streams and highly  
volatile market conditions. As a result, there is a growing need for intelligent, adaptive, and real-time systems  
that can effectively manage stock levels while accurately predicting market prices. Stock management and  
price prediction are inherently interconnected processes. Accurate price forecasts influence purchasing,  
storage, and replenishment decisions, while efficient stock management ensures product availability and cost  
optimization. Conventional statistical methods such as moving averages, linear regression, and autoregressive  
models have been widely used for forecasting; however, these techniques often struggle to capture nonlinear  
relationships and complex temporal patterns present in real-world market data. Moreover, they typically  
require assumptions about data distribution and stationarity, limiting their effectiveness in dynamic  
environments. With the increasing availability of high-frequency data generated from transactions, sensors,  
and online platforms, advanced analytical techniques are required to extract meaningful insights and support  
real-time decision-making. Artificial Intelligence (AI) techniques, particularly machine learning and deep  
www.ijltemas.in  
Page 1219  
INTERNATIONAL JOURNAL OF LATEST TECHNOLOGY IN ENGINEERING,  
MANAGEMENT & APPLIED SCIENCE (IJLTEMAS)  
ISSN 2278-2540 | DOI: 10.51583/IJLTEMAS | Volume XIV, Issue XII, December 2025  
learning, have emerged as powerful tools for addressing the limitations of traditional forecasting and inventory  
management approaches. Machine learning models such as support vector machines, decision trees, and  
ensemble methods are capable of learning complex relationships from historical data, while deep learning  
architectures such as recurrent neural networks (RNNs) and long short-term memory (LSTM) networks are  
well-suited for time-series prediction tasks. These models have demonstrated promising results in stock price  
prediction and demand forecasting due to their ability to handle large datasets and nonlinear dependencies.  
However, relying on a single AI model often leads to challenges such as overfitting, reduced generalization,  
and sensitivity to noise in real-time data streams.  
Real-Time Hybrid Adaptive Forecasting Stock Management and Market Price Prediction  
To overcome these challenges, hybrid AI models have gained increasing attention in recent years. Hybrid  
approaches combine the strengths of multiple AI techniques to achieve improved accuracy, robustness, and  
adaptability shown in fig. 1. By integrating machine learning and deep learning models within a unified  
framework, hybrid systems can effectively capture both short-term market fluctuations and long-term trends.  
Such models are particularly suitable for real-time stock management and market price prediction, where  
decisions must be made continuously based on evolving data. Hybrid AI frameworks also enable the  
incorporation of feature engineering, model fusion, and optimization strategies, enhancing overall system  
performance and reliability. Real-time stock management systems further require seamless integration of  
predictive models with decision-support mechanisms. Accurate price and demand predictions alone are  
insufficient unless they are translated into actionable insights for inventory control. Intelligent stock  
management must dynamically adjust reorder levels, safety stock, and replenishment schedules to minimize  
operational costs while maintaining service quality. Hybrid AI-driven systems offer the capability to automate  
these decisions by continuously learning from new data and updating predictions in real time. This adaptive  
behavior is crucial for businesses operating in fast-changing markets, where delays or inaccuracies in decision-  
making can result in significant financial losses. Traditional forecasting and inventory management approaches  
such as moving averages, linear regression, and autoregressive models have been widely used for demand and  
price prediction; however, these methods often assume data stationarity and linear relationships, which limits  
their effectiveness in highly volatile and dynamic market environments. Furthermore, conventional stock  
management systems rely on static reorder thresholds and historical averages, making them unsuitable for real-  
time decision-making under fluctuating demand and price uncertainty. Recent studies have highlighted that no  
single predictive model consistently performs well across varying market conditions, thereby motivating the  
adoption of hybrid and adaptive AI-based forecasting frameworks . In this context, fig. 1. shows the present  
study focuses on the development of a hybrid AI-based framework for real-time stock management and market  
price prediction. The proposed approach aims to integrate advanced predictive analytics with inventory  
optimization to support intelligent and data-driven decision-making. By leveraging real-time data, hybrid  
modeling techniques, and performance evaluation metrics, the study seeks to demonstrate the effectiveness of  
the proposed system in improving forecasting accuracy and inventory efficiency.  
www.ijltemas.in  
Page 1220  
INTERNATIONAL JOURNAL OF LATEST TECHNOLOGY IN ENGINEERING,  
MANAGEMENT & APPLIED SCIENCE (IJLTEMAS)  
ISSN 2278-2540 | DOI: 10.51583/IJLTEMAS | Volume XIV, Issue XII, December 2025  
LITERATURE REVIEW  
Recent advancements in artificial intelligence have significantly influenced stock market prediction and  
financial decision-support systems. Chachra and Bawa [1] investigated the application of Long Short-Term  
Memory (LSTM) networks trained on historical market data for real-time stock price prediction. Their work  
demonstrated the effectiveness of deep learning models in capturing temporal dependencies in financial time-  
series data; however, the study primarily focused on price prediction and did not address stock management or  
inventory-related decision-making. Alagdeve et al. [2] explored stock price prediction using dual analysis of  
candlestick chart patterns, emphasizing technical indicators derived from historical price movements. While the  
approach improved short-term prediction accuracy, it relied heavily on pattern recognition and lacked  
adaptability to real-time market changes. Additionally, the study did not integrate inventory or operational  
management aspects into the prediction framework. The influence of external textual data on stock price  
movement has been examined by Surulivel et al. [3], who applied natural language processing (NLP)  
techniques for sentiment analysis using news and social media data. Their results indicated that sentiment-based  
features can enhance prediction performance; however, sentiment-driven models often suffer from noise and  
uncertainty and require integration with numerical market data for improved robustness. Similarly, Sharma et  
al. [4] proposed a machine learning framework that integrates market news with stock prices to optimize  
prediction accuracy. Although their approach improved forecasting results, it increased computational  
complexity and was not designed for real-time stock management applications. Yazhinian et al. [5]  
introduced ProStock, a professional stock market navigation and analysis suite aimed at assisting investors  
through analytical tools and visualization techniques. While the system provided valuable insights for decision-  
making, it functioned primarily as an analytical platform and lacked automated prediction and inventory  
optimization capabilities. Joseph et al. [6] focused on stock market analysis and portfolio management,  
emphasizing risk diversification and asset allocation strategies. Their work contributed to investment decision  
support but did not address real-time stock prediction or operational stock control mechanisms. Inani et al. [7]  
presented a bibliometric analysis of deep learning applications in stock market forecasting, highlighting the  
increasing dominance of neural networks such as LSTM, CNN, and hybrid architectures. The study identified  
performance improvements achieved through deep learning but also pointed out challenges related to  
overfitting, interpretability, and computational cost. These challenges indicate the need for hybrid and  
optimized approaches to balance accuracy and efficiency. The application of AI beyond financial trading has  
been explored by Singhal et al. [8], who proposed a smart retail framework utilizing machine learning for  
demand prediction, pricing strategy, and inventory management. Their study demonstrated the benefits of  
predictive analytics in retail operations; however, the system focused more on demand estimation and pricing  
rather than real-time market price prediction. Ahmed et al. [9] evaluated multiple machine learning models for  
financial market prediction and concluded that no single model consistently outperforms others across different  
market conditions, reinforcing the motivation for hybrid modeling approaches. Das et al. [10] proposed an  
improved forecasting model using machine learning techniques to enhance prediction accuracy. Although the  
study reported performance gains over traditional methods, it relied on isolated models and lacked a real-time  
adaptive learning mechanism. Roy et al. [11] extended deep learning techniques to support robo-advisors for  
mutual fund and stock price prediction, demonstrating the growing role of AI in automated financial advisory  
systems. However, the work focused on investment guidance rather than stock management optimization.  
Finally, Fairuzzaky et al. [12] investigated sentiment analysis–based prediction of stock price movements,  
highlighting the relevance of investor sentiment in financial forecasting. Despite its effectiveness, sentiment  
analysis alone was found insufficient for reliable real-time prediction without integration with numerical and  
operational data. From the reviewed literature, it is evident that most existing studies focus on isolated  
prediction models or analytical tools without integrating real-time learning and inventory optimization. While  
deep learning models demonstrate improved forecasting accuracy, they often suffer from high computational  
cost and lack adaptability to sudden market changes. Furthermore, limited attention has been given to  
combining predictive intelligence with operational stock management. These gaps motivate the proposed RT-  
HAF model, which integrates hybrid prediction, real-time adaptability, and inventory decision support within a  
unified framework.  
www.ijltemas.in  
Page 1221  
INTERNATIONAL JOURNAL OF LATEST TECHNOLOGY IN ENGINEERING,  
MANAGEMENT & APPLIED SCIENCE (IJLTEMAS)  
ISSN 2278-2540 | DOI: 10.51583/IJLTEMAS | Volume XIV, Issue XII, December 2025  
PROPOSED METHODOLOGY  
The proposed methodology presents a hybrid artificial intelligence (AI)–based framework for real-time stock  
management and market price prediction, designed to integrate data acquisition, predictive modeling, and  
inventory decision support within a unified architecture. The overall workflow of the system is illustrated  
through sequential stages, ensuring scalability, adaptability, and real-time responsiveness in dynamic market  
environments.  
1. System Architecture: The proposed system architecture consists of five major modules: data collection,  
data preprocessing, hybrid AI-based prediction, inventory optimization, and decision support. Real-time and  
historical data are continuously collected from multiple sources, including transactional databases, historical  
stock price repositories, and market indicators. These data streams are stored in a centralized repository that  
supports real-time processing and model updates. The modular design allows seamless integration of  
predictive models with stock management functions, enabling real-time decision-making.  
2. Data Acquisition and Preprocessing: Data acquisition involves collecting structured and time-stamped  
data such as sales transactions, stock levels, historical prices, and market indicators. To ensure data quality,  
preprocessing techniques including missing value handling, noise removal, outlier detection, and normalization  
are applied. Feature engineering is performed to extract relevant attributes such as moving averages, price  
volatility, demand trends, and seasonality indicators. The processed dataset is then divided into training,  
validation, and testing subsets to support robust model evaluation.  
3. Hybrid AI-Based Prediction Model: The core of the proposed methodology is the hybrid AI prediction  
module, which combines machine learning and deep learning models to improve forecasting accuracy.  
Machine learning models are employed to capture linear relationships and short-term patterns, while deep  
learning models are used to model nonlinear dependencies and long-term temporal trends in stock prices and  
demand data. Model fusion is achieved through ensemble techniques, where predictions from individual  
models are aggregated using weighted averaging or stacking strategies. This hybrid approach reduces model  
bias, enhances generalization, and improves robustness against data variability.  
4. Model Fusion and Optimization Strategy: In the proposed RT-HAF model, predictions generated by  
individual machine learning and deep learning models are combined using an ensemble-based fusion strategy.  
Machine learning models are responsible for capturing short-term and linear market behaviors, while deep  
learning models capture long-term and nonlinear temporal dependencies. The final prediction is computed as a  
weighted aggregation of individual model outputs, where the optimal weights are determined using validation  
error minimization. This fusion mechanism improves robustness, reduces model bias, and enhances overall  
predictive accuracy compared to standalone models.  
5. Real-Time Model Updating Mechanism: To support real-time operation, the proposed system incorporates  
an adaptive learning mechanism that updates model parameters as new data become available. Sliding window  
and incremental learning techniques are employed to ensure that the predictive models remain responsive to  
recent market changes. This continuous learning capability enables the system to handle concept drift and  
sudden market fluctuations effectively, thereby maintaining prediction reliability over time.  
6. Algorithmic Workflow of the Proposed RT-HAF Model: The algorithmic workflow of the proposed RT-  
HAF model begins with real-time data acquisition and preprocessing, followed by feature extraction and  
normalization. Machine learning and deep learning models are trained independently using historical and  
streaming data. The individual model predictions are then fused using an ensemble mechanism to generate  
final forecasts. A sliding window–based incremental learning strategy is applied to update model parameters  
continuously as new data arrive. The final predictions are supplied to the inventory optimization module to  
support real-time stock management decisions.  
6. Inventory Optimization and Decision Support: The predicted stock prices and demand values are utilized  
by an inventory optimization module to support intelligent stock management. Optimization rules are applied  
to determine reorder points, safety stock levels, and replenishment quantities. The decision-support system  
www.ijltemas.in  
Page 1222  
INTERNATIONAL JOURNAL OF LATEST TECHNOLOGY IN ENGINEERING,  
MANAGEMENT & APPLIED SCIENCE (IJLTEMAS)  
ISSN 2278-2540 | DOI: 10.51583/IJLTEMAS | Volume XIV, Issue XII, December 2025  
translates predictive insights into actionable recommendations, such as optimal ordering schedules and  
inventory adjustments. This integration ensures reduced stockouts, minimized overstocking, and improved  
operational efficiency.  
7. Performance Evaluation Metrics: The effectiveness of the proposed methodology is evaluated using  
standard performance metrics such as Mean Absolute Error (MAE), Root Mean Square Error (RMSE), and  
inventory efficiency indicators. Comparative analysis is conducted against standalone predictive models to  
demonstrate the superiority of the hybrid AI approach in terms of accuracy, adaptability, and real-time  
performance.  
RESULT & ANALYSIS  
This section presents the experimental results and performance analysis of the proposed Hybrid AI Model for  
Real-Time Stock Management and Market Price Prediction. The evaluation focuses on forecasting accuracy,  
real-time adaptability, and inventory optimization effectiveness using multiple real-world datasets.  
1. Dataset Description & System Requirements: The experimental study utilizes multiple real-world  
datasets to validate the proposed methodology. The primary dataset consists of historical stock market price  
data obtained from publicly available financial repositories, comprising daily records of opening price, closing  
price, highest price, lowest price, and trading volume over a multi-year period. This dataset captures diverse  
market conditions, including stable, volatile, bullish, and bearish phases. In addition, a sales and inventory  
management dataset is employed to evaluate stock control performance, containing attributes such as daily  
sales volume, current inventory levels, reorder quantities, lead time, and demand variability. To enhance  
predictive accuracy, additional market indicator features, including simple moving averages, exponential  
moving averages, price volatility, and trend indicators, are derived from the raw datasets. All datasets are  
preprocessed using data cleaning, missing value handling, and Min–Max normalization to ensure uniformity  
and improved model convergence. The experiments are conducted on a system equipped with an Intel Core i7  
processor, 16 GB RAM, and a minimum of 4 GB GPU memory, running a 64-bit operating system. The  
proposed framework is implemented using Python with machine learning and deep learning libraries, ensuring  
efficient real-time data processing, model training, and performance evaluation. The historical stock price  
datasets used in this study are obtained from publicly available financial data repositories such as Yahoo  
Finance and Kaggle, which provide open-access time-series market data for research purposes. The inventory  
and sales datasets are sourced from publicly available retail demand datasets commonly used in supply chain  
analytics studies. These datasets ensure transparency, reproducibility, and consistency with prior research in  
stock prediction and inventory optimization.  
2. Comparative Prediction Performance: The proposed hybrid AI model is compared with standalone  
models including Linear Regression (LR), Support Vector Machine (SVM), and LSTM.  
Prediction Accuracy Comparison  
Model  
Linear Regression  
SVM  
MAE  
3.82  
RMSE  
4.96  
3.84  
3.12  
2.41  
2.91  
2.35  
1.78  
LSTM  
Proposed Hybrid AI Model  
TABLE I. shows the hybrid AI model achieves the lowest MAE and RMSE, demonstrating superior prediction  
accuracy. The integration of machine learning and deep learning enables the model to capture both short-term  
fluctuations and long-term trends effectively.  
www.ijltemas.in  
Page 1223  
INTERNATIONAL JOURNAL OF LATEST TECHNOLOGY IN ENGINEERING,  
MANAGEMENT & APPLIED SCIENCE (IJLTEMAS)  
ISSN 2278-2540 | DOI: 10.51583/IJLTEMAS | Volume XIV, Issue XII, December 2025  
Comparative Analysis of Model Prediction Errors  
Fig. 2. shows comparing prediction accuracy of four models like Linear Regression, SVM, LSTM, and  
Proposed Hybrid AI Model using MAE and RMSE metrics. The Proposed Hybrid AI Model shows the lowest  
error values (MAE 1.78, RMSE 2.41), followed by LSTM, SVM, and Linear Regression, indicating improved  
accuracy across successive models.  
3. Real-Time Adaptability Analysis: To assess real-time performance, models were evaluated using a sliding  
window approach with continuous data updates.  
Prediction of Real-Time Adaptability  
Model  
Update Time (ms)  
Accuracy Retention (%)  
LR  
45  
78.2  
82.5  
86.9  
91.4  
SVM  
LSTM  
62  
110  
85  
Proposed Hybrid AI Model  
TABLE II. illustrates the deep learning models require higher computation time, the proposed hybrid  
framework maintains a balance between responsiveness and accuracy, making it suitable for real-time  
applications.  
Real-Time Model Adaptability Performance Comparison  
Fig. 3. illustrates the real-time adaptability of four models such as LR, SVM, LSTM, and the Proposed Hybrid  
AI Model based on update time (milliseconds) and accuracy retention (percentage). LR shows the lowest  
www.ijltemas.in  
Page 1224  
INTERNATIONAL JOURNAL OF LATEST TECHNOLOGY IN ENGINEERING,  
MANAGEMENT & APPLIED SCIENCE (IJLTEMAS)  
ISSN 2278-2540 | DOI: 10.51583/IJLTEMAS | Volume XIV, Issue XII, December 2025  
update time but lower accuracy retention, while LSTM has the highest update time. The Proposed Hybrid AI  
Model achieves the highest accuracy retention (91.4%) with moderate update time, indicating balanced real-  
time adaptability and performance.  
4. Inventory Optimization Results: The impact of accurate prediction on inventory performance is evaluated  
using service level and cost reduction metrics.  
Comparative Inventory Performance of Traditional and Proposed Hybrid AI Systems  
Traditional  
System  
Proposed Hybrid AI  
System  
Metric  
Stockout Rate (%)  
12.6  
4.3  
Overstock Level (%)  
18.9  
85.4  
—
7.8  
Inventory Service Level (%)  
Inventory Cost Reduction (%)  
96.2  
21.5  
The proposed system significantly reduces stockouts and overstocking while improving service levels.  
Accurate demand and price predictions enable optimized reorder decisions and efficient stock utilization.  
TABLE III. compares inventory performance metrics between a traditional system and a proposed hybrid AI  
system. It shows that the hybrid AI system significantly reduces stockout rate and overstock level, improves  
inventory service level, and achieves notable inventory cost reduction compared to the traditional approach.  
Comparative Inventory Performance of Traditional and AI-Based Systems  
Fig. 4. shows comparing inventory performance metrics between a Traditional System and a Proposed Hybrid  
AI System. Metrics include stockout rate, overstock level, inventory service level, and inventory cost  
reduction. The Proposed Hybrid AI System shows significantly lower stockout and overstock rates, a higher  
inventory service level, and achieves notable inventory cost reduction, whereas cost reduction is not reported  
for the Traditional System.  
The experimental results confirm that the hybrid AI approach outperforms individual predictive models in both  
forecasting accuracy and inventory efficiency. The ensemble-based fusion strategy enhances robustness  
against noisy and volatile market data. Moreover, the real-time learning mechanism ensures adaptability to  
sudden market changes, making the system suitable for dynamic business environments. The integration of  
prediction and decision-support modules bridges the gap between analytics and operational execution,  
providing a practical and scalable solution for real-time stock management.  
www.ijltemas.in  
Page 1225  
INTERNATIONAL JOURNAL OF LATEST TECHNOLOGY IN ENGINEERING,  
MANAGEMENT & APPLIED SCIENCE (IJLTEMAS)  
ISSN 2278-2540 | DOI: 10.51583/IJLTEMAS | Volume XIV, Issue XII, December 2025  
CONCLUSION  
This paper presented a hybrid artificial intelligence–based framework for real-time stock management and  
market price prediction, aimed at improving forecasting accuracy and inventory decision-making in dynamic  
market environments. By integrating multiple AI models, the proposed system effectively captured both linear  
and nonlinear patterns in stock price and demand data, leading to improved prediction accuracy compared to  
traditional and single-model approaches. The experimental results demonstrated a significant reduction in  
stockout rates and overstock levels, along with an improvement in inventory service levels, confirming the  
practical effectiveness of the proposed approach. The simplicity, adaptability, and real-time capability of the  
hybrid model make it suitable for applications in retail, supply chain management, and financial markets. As  
future work, the framework can be extended by incorporating additional external factors such as market  
sentiment, news analytics, and macroeconomic indicators, as well as exploring reinforcement learning and  
blockchain-based integration to further enhance decision automation, transparency, and scalability in real-  
world stock management systems. Quantitatively, the proposed RT-HAF model achieved lower prediction  
errors (MAE = 1.78, RMSE = 2.41) and enabled an inventory cost reduction of approximately 21.5%  
compared to traditional systems.  
REFERENCES  
1. P. Chachra and D. S. Bawa, "Training LSTM on Historical Market Data for Predicting Stock Prices on  
Real-Time Basis," 2025 12th International Conference on Computing for Sustainable Global  
Development  
(INDIACom),  
Delhi,  
India,  
2025,  
pp.  
1-6,  
doi:  
10.23919/INDIACom66777.2025.11115302.  
2. V. Alagdeve, P. Karule and P. Palsodkar, "Stock Prediction Using Dual Analysis of Candlestick Chart  
Pattern," 2024 2nd International Conference on Emerging Trends in Engineering and Medical Sciences  
(ICETEMS), Nagpur, India, 2024, pp. 497-502, doi: 10.1109/ICETEMS64039.2024.10965039.  
3. S. T. Surulivel, M. Ashfaq Sultan, K. Ajithkumar and S. Selvabaskar, "Sentiment Analysis in Financial  
Markets: Using NLP to Predict Stock Price from News and Social Media Data," 2025 International  
Conference on Visual Analytics and Data Visualization (ICVADV), Tirunelveli, India, 2025, pp. 25-  
29, doi: 10.1109/ICVADV63329.2025.10961060.  
4. A. Sharma, S. Kumar, C. Shyamsunder, B. Devender, V. S. Nishok and K. S. Sidhu, "Optimizing Stock  
Market Prediction Through Integration of Market News and Stock Prices with Machine Learning  
Models," 2024 IEEE 9th International Conference on Engineering Technologies and Applied Sciences  
(ICETAS), Bahrain, Bahrain, 2024, pp. 1-6, doi: 10.1109/ICETAS62372.2024.11120199.  
5. S. Yazhinian, Y. Hanumanthu, U. A. Srinivas, P. Srikrishna, M. L. Prasanna and K. T. Kumar,  
"ProStock: Professional Stock Market Navigation and Analysis Suite," 2025 International Conference  
on Advanced Computing Technologies (ICoACT), Sivalasi, India, 2025, pp. 1-5, doi:  
10.1109/ICoACT63339.2025.11004817.  
6. M. Joseph, D. P. Kumar and J. K. Keerthana, "Stock Market Analysis and Portfolio  
Management," 2024 11th International Conference on Reliability, Infocom Technologies and  
Optimization (Trends and Future Directions) (ICRITO), Noida, India, 2024, pp. 1-6, doi:  
10.1109/ICRITO61523.2024.10522179.  
7. S. K. Inani, H. Pradhan, G. Nagpal, V. Pagaria and X. -H. Nghiem, "Applications of Deep Learning  
Techniques in Stock Market Forecasting: A Bibliometric Analysis," 2023 4th International Conference  
on Computation, Automation and Knowledge Management (ICCAKM), Dubai, United Arab Emirates,  
2023, pp. 01-06, doi: 10.1109/ICCAKM58659.2023.10449584.  
8. K. Singhal, V. Singh and A. Kaul, "Smart Retail: Utilizing Machine Learning for Demand Prediction,  
Price Strategy, and Inventory Management," 2024 IEEE 16th International Conference on  
Computational Intelligence and Communication Networks (CICN), Indore, India, 2024, pp. 485-489,  
doi: 10.1109/CICN63059.2024.10847534.  
9. K. R. Ahmed, M. M. H. Sizan, A. Chouksey, M. A. A. Montaser, M. R. Islam and M. S. Islam, "AI for  
Financial Market Prediction: Machine Learning Models for Stock Price Forecasting," 2025  
International Conference on Networks and Cryptology (NETCRYPT), New Delhi, India, 2025, pp.  
739-744, doi: 10.1109/NETCRYPT65877.2025.11102142.  
www.ijltemas.in  
Page 1226  
INTERNATIONAL JOURNAL OF LATEST TECHNOLOGY IN ENGINEERING,  
MANAGEMENT & APPLIED SCIENCE (IJLTEMAS)  
ISSN 2278-2540 | DOI: 10.51583/IJLTEMAS | Volume XIV, Issue XII, December 2025  
10. S. Das, D. K. Mohanta, G. K. Mahato and M. Siddique, "Improved Forecasting Model for Stock  
Market Prediction Using Machine Learning Techniques," 2024 2nd International Conference on Signal  
Processing, Communication, Power and Embedded System (SCOPES), Paralakhemundi Campus,  
Centurion University of Technology and Management, Odisha., India, 2024, pp. 1-6,  
doi:10.1109/SCOPES64467.2024.10991322.  
11. S. K. Roy, N. Narayan Das, Savita, N. Kumar Bhasin, S. Kadyan and A. R. K, "Empowering Robo-  
Advisors: Data-Driven Mutual Fund and Stock Market Price Prediction with Deep Learning  
Techniques," 2024 Third International Conference on Electrical, Electronics, Information and  
Communication  
Technologies  
(ICEEICT),  
Trichirappalli,  
India,  
2024,  
pp.  
1-6,  
doi:  
10.1109/ICEEICT61591.2024.10718557.  
12. M. Fairuzzaky, R. Jayadi and E. Miranda, "Utilization of Sentiment Analysis to Predict Stock Price  
Movement," 2025 International Conference on Smart Computing, IoT and Machine Learning (SIML),  
Surakarta, Indonesia, 2025, pp. 1-6, doi: 10.1109/SIML65326.2025.11081041.  
www.ijltemas.in  
Page 1227