Informed Decision Making with Sentiment Analysis and Business Intelligence

Used for final exam assignments - distributed big data courses

Transcript

1. Enhancing Business Intelligence: Integrating Sentiment Analysis for Informed Decision-Making
2. Background – Why need sentiment analysis • Information of user satisfaction • Stay current with the market trends • Getting know about perception of user when use our products
3. Background – What company wants • To stay ahead with the other competitors • Informed decision-making • Improve business performance while maintaining the quality of products
4. How to overcome all of these things?
5. Bridging gap Business Intelligence Machine Learning Potential Outcomes
6. Potential Outcomes • Comprehensive understanding of user behaviors • Adaptability • Minimize faults when making decisions
7. Conceptualization • SVM-based model ◦ Great to classify user feedback • Datasets comes from AppStore, PlayStore review or others social media • Sentiment classification will be divided into 3: positive, negative and neutral
8. High-level Concept
9. Evaluation & Results • F1-Score, Recall, Precision, Accuracy • The results would be visualized on Business Intelligence dashboard ◦ Visualization from sentiment analysis results ◦ Automation classification based on sentiment analysis results ◦ Customizable dashboards based on certain time
11. Challenge(s) & Limitation(s) • Biased Classification • Diversification from User Reviews • Outliers • Resources (manpower, time, capital)
12. Another considerations • Scalability • Security • Monitoring
13. What’s next?
14. Enrich with the new feature for the BI dashboard Monitoring, Security, Scalability Improve the ML model for getting the better results Continuous Improvements
15. Conclusion • One of the key strategic to drive the business operation for organization • The integration between ML application into BI dashboard will give a comprehensive and holistic point of view • Despite enormous challenges, still have a room for improvements