HCI-SLR Systematic Literature Review Assignment

Literature Review based on MAchine learning

1. LITERATURE REVIEW (DOMIAN KNOWLEDGE) Presented by: Rashid Rehmat PHD-Computer Science Department of Computer Science, Superior University Lahore

2. Project Topic “Prediction or Diagnosis of Heart Diseases for Diabetic Patients using Supervised Machine Learning Approach”

3. Introduction According to estimates from the World Health Organization (WHO), more than 17.9 million individuals pass away each year as a result of cardiovascular diseases. [1]. When one or more coronary arteries are clogged, the heart does not receive enough blood, which results in a heart attack. When cholesterol-rich fat deposits accumulate and form plaques, which are blood clots, the obstruction occurs. The main risk factors for developing these clots are an unhealthy lifestyle, using tobacco or alcohol, being overweight or obese and having high blood pressure or raised blood sugar levels. Several machine learning models have been proposed in different studies, which will be applied to identify patients who are more likely to experience a heart attack. Studies are still being conducted to improve and enhance the prediction techniques through machine learning .

4. Literature Review (SYSTEMATIC LITERATURE REVIEW) A systematic literature review (SLR) is an  independent academic method  that aims to identify and evaluate all relevant literature on a topic in order to derive conclusions about the question under consideration . " Systematic reviews are undertaken to clarify the state of existing research and the implications that should be drawn from this." ( Feak & Swales, 2009, p. 3)

5. Literature Review (SLR#1) “ Heart Attack Prediction using Machine learning: A Comprehensive SLR and Bibliometric Analysis ” Journal of Theoretical and Applied Information Technology 15th March 2024. Vol.102. No 5 www.jatit.org

6. Literature Review (Introduction) The review focuses on the use of machine learning (ML) algorithms for predicting heart attacks. The SLR covered studies from 2017 to 2021, aiming to evaluate and summarize the methodologies and effectiveness of ML techniques in this area.

7. Literature Review (METHODOLOGY)

8. RESEARCH QUESTIONS RQ1: Who are the most productive authors in research on Heart Attack Prediction using Machine Learning? RQ2 : What methodologies are being used for Machine Learning development? RQ3 : What are the criteria for measuring the overall effectiveness of Heart Attack Prediction using Machine Learning? RQ4 : What are the most commonly used algorithms for Machine Learning development? RQ5 : What are the most used topics in research on the prediction of heart attacks using Machine Learning?

9. Search Sources and Search Strategies The SLR process involved searching databases like Taylor and Francis, IEEE Xplore , ARDI, ACM Digital Library, ProQuest , Wiley Online Library, and Microsoft Academic, resulting in 3,525 initial articles, which were narrowed down to 82 relevant studies.

10. Search Sources and Search Strategies

11. Studies Identified

12. Methodology The exclusion criteria's used are detailed below: CE1: Articles are older than 5 years. CE2: Articles are not written in English. CE3: Articles are not published in conferences or journals . CE4: Titles and keywords of articles are not very suitable . CE5: Articles are not unique. CE6: The abstract of articles is not very relevant.

13. Study Selection

14. Review Findings The review identified common themes across different studies, highlighting the most effective ML algorithms for heart attack prediction . Various methodologies, including classification and regression techniques, were found to significantly enhance prediction accuracy . The findings were synthesized to address specific research questions

15. Review Findings The information above shows that the countries with the highest number of studies are India with 24 (29.27%), the United Kingdom with 18 (21.95%), and the United States with 14 (17.07%).

16. Conclusion The study concludes that while there are various ML approaches heart attack prediction, consistent studies are aimed at improving prediction accuracy . This review provides a comprehensive guide for researchers, highlighting the current state of ML applications in heart attack prediction and identifying areas for future improvement .

17. Literature Review (SLR#2) “ What can machines learn about heart failure? A systematic literature review ” International Journal of Data Science and Analytics Volume 13, pages 163–183, (2022)

18. Literature Review (SLR-2: Introduction) The paper systematically reviews the use of data science and ML in HF datasets to identify common problems, assess model performance and utility of ML techniques. The review includes 81 studies, focusing on detection, readmission prediction, mortality prediction, classification, and clustering of HF cohorts .

19. Literature Review (SLR-2: Introduction) This systematic literature review is comprised of eight sections Section 1: Introduction, outlines the rationale for the review. Section 2: provides a summary of the previous literature reviews Section 3: provides the search criteria, and a and inclusion-exclusion criteria. Section 4: Results Section 5: Discussion and Guidelines Section 6 : Gaps and Research Opportunities Section 7 : Conclusions

20. Search Sources and Search Strategies SCOPUS , ProQuest and MEDLINE Ovid databases were searched using the following terms - search terms included (“heart failure” OR cardiomyopathy/ ies OR “cardiac oedema ” OR “paroxysmal dyspnoea ”) AND (“machine learning” OR “data mining” OR “data analytics” OR “data science”)

21. Inclusion and exclusion criteria

22. Common HF problems addressed by ML Detection of HF Prediction of hospital readmission and mortality Classification and clustering of HF cohorts

23. Common HF problems addressed by ML This Figure presents the common HF problems and most commonly used ML algorithms in addressing these specific HF domain problems..

24. Gaps and Research Opportunities Several gaps and future research opportunities are identified: Clinical Pathways: Integration of ML into clinical decision-making processes. Access to Modern HF Databases: Utilization of contemporary and diverse datasets. Validation of Algorithms: Need for prospective validation in clinical settings. Future ML Methods: Exploration of advanced ML techniques. Collaboration: Enhanced collaboration between data curators and clinicians. Transparency: Improved transparency and reporting in ML trials. Randomized Controlled Trials (RCTs): Urgent need for RCTs to test ML models in real-world clinical environments .

25. Conclusion The review concludes that while ML has shown potential in addressing HF-related problems, significant work is needed to validate these models in clinical practice. This systematic literature review provides a critical evaluation of current ML applications in HF and outlines essential steps for advancing the field .

26. Literature Review (List of Articles Reviewed) Article #1: Clinical applications of machine learning in cardiovascular disease and its relevance to cardiac imaging Article #2: Machine learning prediction in cardiovascular diseases: a meta-analysis Article #3: Using Machine Learning Algorithms in Cardiovascular Disease Risk Evaluation Article #4: Identification of Cardiovascular Diseases Using Machine Learning Article #5: Prediction of Cardiovascular Disease Using Machine Learning Algorithms (IEEE)

27. Literature Review (List of Articles Reviewed) Article #6: A data-driven approach to predicting diabetes and cardiovascular disease with machine learning Article #7: Machine learning techniques for classification of diabetes and cardiovascular diseases Article #8: Machine Learning and the Future of Cardiovascular Care Article #9: Applying Machine Learning Methods in Diagnosing Heart Disease for Diabetic Patients Article #10: Machine learning for diabetes clinical decision support: a review

28. ML Models Used Artificial Neural Networks (ANNs ) k-Nearest Neighbors Naïve Bayesian network (BN) Logistic regression S upport vector machines (SVM) R andom forest Gradient boosting

29. THANK YOU