mini_proj[1][1] 1-1.pptxdvdsgdsgdsdsdsfdfdf

dsfafdsfdsfsa

1. Date: 0 8 .09.2025 An Autonomous Institution Biotech, Chemical,ECE &IT Department of Artificial Intelligence and Data Science HYBRID STOCK PRICE PREDICTION WITH SENTIMENT ANALYSIS AND EXPLAINABLE AI PRESENTED BY VH12918 Blessy Miraculine P.D - 113023243018 - V Sem VH12935 Janani B.S              - 113023243035 - V Sem VH12937 Jayasri N  - 113023243037 - V Sem Supervisor: Mrs.Harini P, Assistant Professor ssor Head of the Department: Mrs.Geetha L ,Assistant Professor 21EE01P-MINI PROJECT

2. PROJECT STRUCTURE The proposed project work consisting of the Chapters as follows. Chapter 1: Introduction Chapter 2 : Objective Chapter 3 : Literature survey Chapter 4: Architecture Diagram Chapter 5 : Proposed methodology Chapter 6: Experiments and Results Chapter 7 : Discussion and Conclusion

3. INTRODUCTION This project presents a hybrid stock prediction system that combines historical market data with sentiment analysis of financial news. Using advanced machine learning and Explainable AI (XAI) techniques, it delivers accurate forecasts while providing clear and transparent insights for smarter, informed investment decisions.

4. OBJECTIVE To build a hybrid stock price prediction system by combining historical market data with sentiment analysis derived from news, social media, and other public sources. To incorporate Explainable AI (XAI) methods to provide transparent and interpretable predictions, supporting investors in making confident and informed decisions. To Implement a multilingual stock prediction system that combines market data and financial news for accurate, transparent global forecasting.

5. LITERATURE SURVEY RESEARCH ARTICLE :01 TITLE OF THE RESEARCH ARTICLE A hybrid model for stock price prediction using machine learning technique with CNN. PROBLEM ADDRESSED / IDENTIFIED The paper addresses the challenge of accurately predicting stock prices influenced by multiple dynamic factors. AIM / OBJECTIVES The study aims to build a CNN–LSTM hybrid model for improved stock price prediction accuracy. NOVELTY / SIGNIFICANCE The approach merges CNN feature extraction with LSTM to boost prediction accuracy. LIMITATION / WEAKNESS The model ignores external factors like news and events that impact stock prices FINDINGS / CONCLUSION The CNN–LSTM hybrid outperformed single models, achieving the highest accuracy (R²≈0.90). AREAS OF IMPROVEMENT / FUTURE ENHANCEMENT Future work includes adding news-based sentiment analysis to incorporate external factors.

6. PROPOSED METHODOLOGY Historical stock data and sentiment information from news, tweets, and forums are collected and preprocessed. Technical indicators are extracted from the price data, while sentiment scores are computed from the textual content, and both are combined into a unified feature set. The integrated features are used to train machine learning models. Explainable AI techniques such as SHAP and DiCE are incorporated. SHAP explains predictions by showing each feature’s contribution, while DiCE provides counterfactual examples, suggesting how small input changes could alter the outcome . The model generates stock price predictions on new data, with interpretability outputs explaining each prediction, ensuring transparency and supporting informed investment decisions. We also integrate multilingual support for broader accessibility .

7. Architecture Diagram

8. Source Code

9. SAMPLE OUTPUT

10. SAMPLE OUTPUT

11. SAMPLE OUTPUT

12. SAMPLE OUTPUT Re-evaluation shows higher negative risk (68%). Positive outlook dropped from 42% to 32%. Final recommendation: Sell due to elevated risk.

13. DISCUSSION AND CONCLUSION Our analysis shows that combining historical stock data with sentiment features improves prediction accuracy compared to using historical data alone. SHAP explanations give clear and consistent insights, though they take slightly more time to compute. Overall, the hybrid model with historical data, sentiment analysis, and SHAP offers a strong balance of accuracy and transparency, making it a more reliable tool for informed investment decisions. It turns complex AI predictions into a clear guide, so even someone without deep stock market knowledge can make smarter investment choices.

14. REFERENCES References: 1. Li, X., Wu, P., & Wang, W. Incorporating stock prices and news sentiments for stock market prediction: A case of Hong Kong. Information Processing & Management, 57(5), 102212. (2023) 2. Malandri , L., Xing, F. Z., Orsenigo , C., Vercellis , C., & Cambria, E. Public mood–driven asset allocation: The importance of financial sentiment in portfolio management. Cognitive Computation, 10(6), 1167–1176. (2020) 3. Dogan, E., & Kaya, B. Deep learning based sentiment analysis and text summarization in social networks. 2019 International Artificial Intelligence and Data Processing Symposium (IDAP). IEEE. (2022). 4. Li, H., & Hu, J. A hybrid deep learning framework for stock price prediction considering the investor sentiment of online forums enhanced by popularity.(2024) 5. Zhang, Y., & Chen, X. A hybrid model integrating deep learning with investor sentiment analysis for stock price prediction. (2021)

15. Thank you *