Page 74 - Kaleidoscope Academic Conference Proceedings 2024

2024 ITU Kaleidoscope Academic Conference

QSVC showed a relatively good performance over the computational velocity, so QML would be more suitable to
Indian PIMA diabetes dataset with an accuracy of 0.74 solve big medical data analysis problem.
and the precision, recall, and F-measure of 0.75,0.73,0.74
respectively (Figure 8). Whereas the VQC has the lowest REFERENCES
accuracy of 0.61 and precision, recall, and F-measure of
0.60,0.61, and 0.59 respectively which poses a lesser value [1] “Diabetes,” https://www.who.int/news-room/
compared with the other classifiers. fact-sheets/detail/diabetes, Accessed: Apr.
05, 2023.
[2] “Diabetes Cure for Type1, Type 2 and Type 3 (LADA),”
International Journal of Diabetes & Metabolic
Disorders, 2018.

[3] “Diabetes: Mody,” https://www.
nationwidechildrens.org/conditions/
diabetes-mody#, Accessed: Apr. 14, 2024.

[4] R.I. Mukhamediev et al., “Review of Artificial
Intelligence and Machine Learning Technologies:
Classification, Restrictions, Opportunities and
Challenges,” Mathematics, 2022.
[5] I.H. Sarker, “Machine Learning: Algorithms,
Figure 8 – Comparative Analysis of VQC and QSVC.
Real-World Applications and Research Directions,” SN
Therefore, it is inferred that VQC is somehow less effective in Computer Science, vol. 2, p. 160, 2021, doi: https:
identifying whether a case of diabetes exists in the given data //doi.org/10.1007/s42979-021-00592-x.
set. These results demonstrate SVM’s superiority in accuracy
and, in particular, the practicality of LR and QSVC for the [6] S. Kaul and Y. Kumar, “Artificial Intelligence-based
diabetes classification problem, while also emphasizing the Learning Techniques for Diabetes Prediction:
necessity for VQC to keep enhancing its accuracy. Challenges and Systematic Review,” SN Computer
Science, vol. 1, p. 322, 2020, doi: https:
//doi.org/10.1007/s42979-020-00337-2.
7. CONCLUSION
[7] T.T. Teoh and Z. Rong, “Classification,” in
This article has presented a comprehensive overview of ML
Artificial Intelligence with Python. Machine
and QML techniques.The nexus of quantum computing with
Learning: Foundations, Methodologies, and
machine learning is known as QML. The study evaluated
Applications, Springer, Singapore, 2022, doi: https:
the potential of QML for the classification of diabetes
//doi.org/10.1007/978-981-16-8615-3_11.
using the PIMA Indian Diabetes Dataset by comparing
quantum-enhanced algorithms (QSVC, VQC) with their [8] K.A. Cyran et al., “Support Vector Machines
classical counterparts (LR, SVM). Future directions is to in Biomedical and Biometrical Applications,” in
Explore more sophisticated QML algorithms for diabetes Emerging Paradigms in Machine Learning. Smart
classification. Investigate the impact of larger datasets Innovation, Systems and Technologies, vol. 13, S.
on QML performance compared to classical methods. Ramanna, L. Jain, and R. Howlett, Eds., Springer,
Analyze the computational efficiency of QML algorithms for Berlin, Heidelberg, 2013, doi: https://doi.org/
real-world applications. 10.1007/978-3-642-28699-5_15.
[9] U. Qamar and M.S. Raza, “Classification,” in Data
8. FUTURE WORK
Science Concepts and Techniques with Applications,
Springer, Cham, 2023, doi: https://doi.org/10.
Further research prospects touching on the use of QML
1007/978-3-031-17442-1_5.
algorithms for diabetes classification are to pursue modified
algorithms with higher effectiveness. As such, it is
[10] M. Bansal, A. Goyal, and A. Choudhary, “A
also necessary to study the work of QML algorithms
comparative analysis of K-Nearest Neighbor, Genetic,
when processing larger datasets than the classical approach.
Support Vector Machine, Decision Tree, and Long
Besides, answers to the questions regarding the computational
Short Term Memory algorithms in machine learning,”
complexity of QML algorithms for real-world applications
Decision Analytics Journal, vol. 3, p. 100071, 2022,
will be obtained. Another significant branch of the subjected
ISSN 2772-6622, doi: https://doi.org/10.1016/
research is considering methodology of applying hybrid
j.dajour.2022.100071.
classical-quantum models combining the advantages of two
approaches. Thus, the application of both approaches [11] S.F. Ahmed, M.S.B. Alam, M. Hassan et
simultaneously could increase predictive accuracy and al., “Deep learning modelling techniques:

– 30 –