ffffffffddffddfICMS - Presentation (1).pptx

dsfsfsdfsfdsfdsfsd

1. Forecasting Stock Index Movement using a Machine Learning Method: Evidence from Pakistan Stock Exchange Saqib Farid- Banking & Finance, Dr. Hassan Murad School of Management. University of Management & Technology, Lahore Contact no: 03334262175 Co-author Quratulain Zafar- Banking & Finance, Dr. Hassan Murad School of Management. University of Management & Technology, Lahore 4 th International Conference on Management Sciences (ICMS – 2024) Lahore Garrison University, Lahore, Pakistan Conference Theme: Cultivating Excellence in the Digital Age

2. Background

3. Theoretical Underpinnings 3 4 th International Conference on Management Sciences (ICMS – 2024) Lahore Garrison University, Lahore, Punjab, Pakistan Conference Theme: Cultivating Excellence in the Digital Age

4. Stock Price Prediction Approaches 4 4 th International Conference on Management Sciences (ICMS – 2024) Lahore Garrison University, Lahore, Punjab, Pakistan Conference Theme: Cultivating Excellence in the Digital Age

5. Application of Machine Learning Method to PSX 5 4 th International Conference on Management Sciences (ICMS – 2024) Lahore Garrison University, Lahore, Punjab, Pakistan Conference Theme: Cultivating Excellence in the Digital Age

6. Methodology 6 4 th International Conference on Management Sciences (ICMS – 2024) Lahore Garrison University, Lahore, Punjab, Pakistan Conference Theme: Cultivating Excellence in the Digital Age

7. Proposed Design 7 4 th International Conference on Management Sciences (ICMS – 2024) Lahore Garrison University, Lahore, Punjab, Pakistan Conference Theme: Cultivating Excellence in the Digital Age

8. Results: 1- Decision Tree Analysis Decision Tree Input Variables   Accuracy Ratio Technical Factors 73.42 % Fundamental Factors 54.31 % Technical + Fundamental Factors 73.52 % 8 4 th International Conference on Management Sciences (ICMS – 2024) Lahore Garrison University, Lahore, Punjab, Pakistan Conference Theme: Cultivating Excellence in the Digital Age Decision tree technical variable importance plot Decision tree fundamental variable importance plot Plot of variable importance with technical-fundamental indicators

9. Results: 2- Random Forest Results Random Forest Input Variables   Accuracy Ratio Technical Factors 75.45 % Fundamental Factors 57.31 % Technical + Fundamental Factors 76.55 % 9 Random forest error rate graph for technical inputs Random forest variable importance graph for technical inputs Random forest variable importance graph for technical inputs

10. Results: 2- Random Forest Results ( Contd ) 10 4 th International Conference on Management Sciences (ICMS – 2024) Lahore Garrison University, Lahore, Punjab, Pakistan Conference Theme: Cultivating Excellence in the Digital Age Random forest variable importance graph for fundamentals Random forest error rate graph for fundamentals Random forest error rate graph for Technical-fundamental indicators Random forest variable importance graph for Technical-fundamental indicators

11. Results: 3- Real Vs Predicted Labels KSE-100 Index Actual Prices and Labels (Selected) & KSE-100 Index Predicted Labels (Selected) 11 Date Price Label 01-01-2021 40711.04 02-01-2021 41486.87 positive 03-01-2021 41544.27 positive 04-01-2021 41908.71 positive 05-01-2021 42523.99 positive 08-01-2021 43112.12 positive 09-01-2021 42814.34 negative Model 2 days ahead 4 days ahead 6 days ahead Decision Tree Positive Positive Positive Random Forest Positive Positive Positive

12. Conclusion 12