VIETNAM NATIONAL UNIVERSITY, HANOI INTERNATIONAL SCHOOL GRADUATION PROJECT FINANCIAL PERFORMANCE ANALYSIS AND AI-BASED STOCK PRICE PREDICTION OF SABECO Tran Quoc Dang Hanoi, 2024 1 VIETNAM NATIONAL UNIVERSITY, HANOI INTERNATIONAL SCHOOL GRADUATION PROJECT FINANCIAL PERFORMANCE ANALYSIS AND AI-BASED STOCK PRICE PREDICTION OF SABECO SUPERVISOR Dr. Ha Manh Hung (Academic title, academic degree, full name) STUDENT Tran Quoc Dang STUDENT ID 20070819 COHORT Machine Learning MAJOR Informatics and Computer Engineering Hanoi, 2024 2 ACKNOWLEDGEMENT I am profoundly thankful to Dr. Ha Manh Hung for his invaluable guidance, support, and insightful contributions throughout the research process. His expertise and encouragement have been crucial to the successful completion of this research paper.
I would also like to express my sincere appreciation to our lecturer for the care and support provided. From the initial stages of ideation to the final phase of completion, his guidance and motivation have been instrumental in helping us overcome challenges along the way. Without his unwavering support, this report would not have been possible. I extend my heartfelt gratitude for his significant contributions and look forward to future collaborations on upcoming projects.
3 TABLE OF CONTENTS ACKNOWLEDGEMENT 3 TABLE OF CONTENTS 4 LIST OF FIGURES 4 LIST OF TABLES 7 LIST OF ABBREVIATIONS 9 ABSTRACT 11 I. Introduction to Financial Analysis 14 2. Stock Price Prediction 14 3. Machine Learning Models in Financial Predictions 14 4.
Previous Studies and Findings 14 5. Summary of Gaps in the Literature 15 III. Software and Tools Used 20 IV. DATA ANALYSIS AND RESULTS 21 1.
Financial Analysis of Sabeco 22 3. Stock Price Prediction 49 4. Model Performance Comparison 65 V. Interpretation of Findings 67 2.
Comparison with Previous Studies 68 4. Implications for Investors 68 5. Limitations of the Study 68 6. Recommendations for Future Research 69 VI.
CONCLUSION AND RECOMMENDATIONS 69 1. Summary of Findings 69 2. Recommendations for Investors 70 3. Future Research Directions 70 REFERENCES 72 4 LIST OF FIGURES Figure 1.
Debt to Equity Ratio comparison between SABECO and 30 other beverage manufacturers Figure 2. Quick Ratio comparison between SABECO and other 32 beverage manufacturers Figure 3. Interest Coverage comparison between SABECO and other 33 beverage manufacturers Figure 4. Number of Days of Payables comparison between SABECO 35 and other beverage manufacturers Figure 5.
Days of Inventory on Hand comparison between SABECO 36 and other beverage manufacturers Figure 6. Days of Sale Outstanding comparison between SABECO 38 and other beverage manufacturers 39 Figure 7. Return on Assets (ROA) comparison between SABECO and other beverage manufacturers Figure 8. Return on Equity (ROE) comparison between SABECO and 41 other beverage manufacturers Figure 9.
Gross Profit Margin comparison between SABECO and 42 other beverage manufacturers Figure 10. Net Profit Margin comparison between SABECO and other 44 beverage manufacturers Figure 11. Price to Book Ratio comparison between SABECO and 45 5 other beverage manufacturers Figure 12. Price to Earning Ratio comparison between SABECO and 47 other beverage manufacturers Figure 13.
30-day Moving Average for SAB 49 Figure 14. 30-day Moving Average for VN-INDEX 50 Figure 15. LSTM Model - VN-INDEX Predicting Stock plot 58 Figure 16. GRU Model - VN-INDEX Predicting Stock plot 58 Figure 17.
Bidirectional LSTM Model - VN-INDEX Predicting Stock 59 plot Figure 18. Bidirectional GRU Model - VN-INDEX Predicting Stock 59 plot Figure 19. LSTM Model - SAB Predicting Stock plot 60 Figure 20. GRU Model - SAB Predicting Stock plot 60 Figure 21.
Bidirectional LSTM Model - SAB Predicting Stock plot 61 Figure 22. Bidirectional GRU Model - SAB Predicting Stock plot 61 Figure 23. VN-INDEX Stock Prices for the next 180 days plot 62 6 Figure 24. SAB Stock Prices for the next 180 days plot 62 LIST OF TABLES Table 1.
Beverage Manufacturing Data dataset 15 Table 2. MA30 Data dataset 16 Table 3. SAB Financial Data dataset 17 Table 4. Statement of Financial Position 21 Table 5.
Statement of Profit or Loss 25 Table 6. Data collected from audited financial statements of BHN, 27 SAB, SCD, SMB, HAD, HAT, THB, VDL companies from 2020-2022 Table 7. Category and Category 2 table for Master data 28 Table 8. Final data to use for analysis of the financial of Beverage 29 Manufacturing industry Table 9.
LSTM Network Architecture table 54 Table 10. GRU Network Architecture table 54 Table 11. Bidirectional LSTM Network Architecture table 55 Table 12. Bidirectional GRU Network Architecture table 55 Table 13.
Model Performance Comparison 63 7 LIST OF ABBREVIATIONS Abbreviation Full Form AI Artificial Intelligence BHN Hanoi Beer Alcohol And Beverage Joint Stock Corporation D/E Debt to Equity Ratio EBIT Earnings Before Interest and Taxes F&B Food and Beverage GRU Gated Recurrent Unit HAD Ha Noi - Hai Duong Beer JSC HAT Ha Noi Beer Trading Joint Stock Company HOSE Ho Chi Minh Stock Exchange HNX Hanoi Stock Exchange LSTM Long Short-Term Memory MAE Mean Absolute Error MSE Mean Squared Error P/B Price to Book Ratio P/E Price to Earnings Ratio 8 ROA Return on Assets ROE Return on Equity SAB Saigon Beer - Alcohol - Beverage Corporation SCD Chuong Duong Beverages Joint Stock Company SMB Sai Gon - Mien Trung Beer JSC SVM Support Vector Machine THB Ha Noi - Thanh Hoa Beer Joint Stock Company VDL Lam Dong Foodstuffs JSC 9 ABSTRACT The accurate prediction of stock prices is a critical task in financial markets, offering significant benefits to investors and financial analysts. This study focuses on the financial analysis and stock price prediction of the Saigon Beer - Alcohol - Beverage Corporation (Sabeco), a prominent player in Vietnam's beverage industry. By leveraging advanced machine learning models, including GRU, bidirectional GRU, LSTM, bidirectional LSTM, this study aims to enhance the accuracy of stock price forecasts. The methodology involves collecting and preprocessing financial data from Sabeco and other relevant companies, developing various predictive models, and evaluating their performance using historical data.
The findings of this study provide valuable insights for investors, highlighting the strengths and limitations of different AI models in predicting stock prices. The results indicate that integrating financial analysis with advanced AI techniques can significantly improve the reliability of stock price predictions, offering a robust tool for investment decision-making. Keywords Stock price prediction, financial analysis, machine learning, GRU, LSTM, bidirectional GRU, bidirectional LSTM, Sabeco. Background of the Study The financial well-being of a company is a crucial determinant of its stock price performance.
In the rapidly evolving financial markets, investors depend heavily on accurate stock price predictions to make informed investment decisions. This study centers on the Saigon Beer - Alcohol - Beverage Corporation (Sabeco), a leading company in Vietnam's beverage industry. By analyzing its financial indicators and employing advanced AI models to forecast its stock prices, this research aims to provide valuable insights for investors. Problem Statement Predicting stock prices is inherently complex due to the volatile and non-linear nature of financial markets.
Traditional statistical methods often fail to capture the intricate patterns in stock price movements. This study aims to address this challenge by utilizing advanced machine learning models to improve the accuracy of stock price predictions for Sabeco. Additionally, state policies like Decree 100, which prohibits driving after drinking, further complicate the prediction of stock prices and must be taken into account due to their significant impact on consumption patterns. Objectives of the Study The specific objectives of this study are: ● To conduct a comprehensive financial analysis of Sabeco.
● To accurately forecast the future stock prices of Sabeco using various AI models, including GRU, bidirectional GRU, LSTM, and bidirectional LSTM. ● To compare the performance of these AI models and identify the most effective model for stock price prediction. ● To consider the impact of state policies like Decree 100 on Sabeco’s financial performance and stock prices. Research Questions ● How do the financial indicators of Sabeco influence its stock price? ● Which AI model provides the most accurate predictions for Sabeco's stock prices? ● How do the predictions of different AI models compare in terms of accuracy and reliability? ● How do state policies like Decree 100 affect the financial performance and stock prices of Sabeco? 5.
Significance of the Study This study enhances the existing body of knowledge by integrating advanced AI models with financial analysis to predict stock prices. The findings can aid investors in making better-informed decisions and can also serve as a reference for future research in financial market predictions. Additionally, considering the impact of state policies like Decree 100 provides a more comprehensive understanding of the factors influencing stock prices. Scope and Limitations The scope of this study is confined to the financial data of Sabeco and selected companies in the Food and Beverage industry from 2020 to 2022.
While the study employs multiple AI models, it does not encompass every possible machine learning technique. Furthermore, the study's predictions are based on historical data, and future market conditions may introduce unforeseen variables. The impact of state policies like Decree 100 is considered qualitatively due to the unavailability of new datasets. Introduction to Financial Analysis Financial analysis involves evaluating a company's financial statements to understand its economic health and performance.
Key financial indicators, such as revenue, profit margins, return on assets (ROA), and the debt-to-equity ratio (D/E), are essential in determining a company's value and growth potential. In the context of stock price prediction, these indicators offer valuable insights into the company's stability and future performance. Stock Price Prediction Stock price prediction is a well-researched area in financial studies. Traditional methods like linear regression and time series analysis have been widely used but often fail to capture the non-linear patterns in stock price movements.
Advanced machine learning techniques, such as neural networks and ensemble methods, provide promising alternatives by learning complex patterns from historical data. Machine Learning Models in Financial Predictions Machine learning models have gained popularity in financial predictions due to their ability to handle large datasets and uncover hidden patterns. Recurrent neural networks (RNN), particularly Long Short-Term Memory (LSTM) networks, are effective for time series prediction due to their memory capabilities. These models play significant roles in classification and clustering tasks within financial data analysis.
Previous Studies and Findings Numerous studies have demonstrated the efficacy of machine learning models in stock price prediction. For example, research has shown that LSTM networks outperform traditional time series models in predicting stock prices due to their ability to capture long-term dependencies. Tree-based models like Random Forest 13 and XGBoost have been successful in feature selection and improving prediction accuracy. However, each model has its limitations, and the choice of model depends on the specific characteristics of the dataset and the prediction task.
Summary of Gaps in the Literature Despite the advancements in machine learning models, there are still gaps in the literature, particularly in comparing the performance of different models on the same dataset. Additionally, integrating financial analysis with advanced AI models remains underexplored. This study aims to fill these gaps by comparing various AI models and providing a comprehensive analysis of their performance in predicting Sabeco's stock prices. Moreover, the impact of state policies like Decree 100 on financial performance is underexplored, and this study aims to address this gap.
Research Design This study utilizes a quantitative research design to analyze financial data and predict stock prices using various machine learning models. The research follows a structured approach that includes data collection, data preparation, model development, and evaluation. Data Collection ● Sources of Data The primary data source for this study is Vietstock, which provides comprehensive financial reports of companies in Vietnam. The study focuses on the financial data of Sabeco and other relevant companies in the Food and Beverage industry from 2020 to 2022.
● Description of Data The datasets include: Beverage Manufacturing Data: Indicators related to accumulated depreciation and other financial metrics for companies such as BHN and SAB. Beverage Manufacturing Data dataset. 15 MA30 Data: VN-INDEX and SAB stock prices over time, along with their 30-day moving averages. MA30 Data dataset.