Obesity Analysis and Prediction with Optimized Supervised Learning Algorithms

Öz

Obesity is a health problem that is the fifth most important cause of death worldwide. In the report published in 2022, the World Health Organization (WHO) emphasized that obesity forms the basis of many diseases and can be stopped by following the necessary measures and policies. Therefore, obesity analysis and prediction applications with machine learning algorithms are important. In this study, prediction models were developed with K-Nearest Neighbor Algorithm (KNN) and Random Forest Algorithms (RF), which are supervised learning algorithms, using data from the UCI machine learning data repository. These models were compared using different statistical evaluation criteria. As a result of the evaluation, the RF model with hyperparameter optimization achieved the best prediction result with an average accuracy of 94%. The study is important because it analyzes and visualizes the factors affecting the prevalence of obesity and predicts its levels with a high success rate.

Anahtar Kelimeler

• Obesity Analysis
• Supervised Machine Learning
• Hyperparameter Optimization
• KNN
• RF

Optimize Edilmiş Denetimli Öğrenme Algoritmaları ile Obezite Analizi ve Tahmini

Öz

Obezite dünya genelinde gerçekleşen ölümlerin en önemli beşinci nedeni olarak karşımız çıkan bir sağlık sorunudur. Dünya Sağlık Örgütü (DSÖ) 2022 yılında yayınladığı raporda obezitenin birçok hastalığın temelinin oluşturduğunu ve gerekli önlemler ve politikalar izlenerek durdurulabileceğini vurgulamıştır. Bu nedenle makine öğrenmesi algoritmaları ile obezite analizi ve tahmin uygulamaları önemlidir. Bu çalışmada, UCI makine öğrenmesi veri havuzundan alınan veriler kullanılarak, denetimli öğrenme algoritmalarından K-En Yakın Komşu Algoritması(KNN) ve Rastgele Ormanlar Algoritmaları(RF) ile tahmin modelleri geliştirilmiştir. Bu modeller farklı istatistiksel değerlendirme kriterleri kullanılarak karşılaştırılmıştır. Değerlendirme sonucunda hiper parametre optimizasyonu gerçekleştirilen RF modeli %94 ortalama accuracy(doğruluk) sonucu ile en iyi tahmin sonucunu elde etmiştir. Çalışma obezite prevalansını etkileyen faktörleri analiz etmesi, görselleştirmesi ve yüksek bir başarı oranı ile seviyelerini tahmin etmesiyle önemlidir.

Anahtar Kelimeler

• Obezite Analizi
• Denetimli Makine Öğrenmesi
• Hiperparametre Optimizasyonu
• KNN
• RF

Kaynakça

1. Bansal, S., Jin, Y. (2023). Heterogeneous Effects of Obesity on Life Expectancy: A Global Perspective. Annual Re-view of Resource Economics, 15; DOI 10.1146/annurev-resource-022823-033521.
2. Bashir, M. B., Abd Latiff, M. S. B., Coulibaly, Y.,Yousif, A. (2016). A survey of grid-based searching techniques for large scale distributed data. Journal of Network and Com-puter Applications, 60, 170-179.
3. Breiman, L. (2001). Random forests. Machine learning, 45: 5-32.
4. Brero, M., Meyer, C. L., Jackson‐Morris, A., Spencer, G., Ludwig‐Borycz, E., Wu, D., Nugent, R. (2023). Invest-ment case for the prevention and reduction of childhood and adolescent overweight and obesity in Mexi-co. Obesity Reviews, 24(9); DOI 10.1111/obr.13595.
5. Caballero, B. (2005). A nutrition paradox underweight and obesity in developing countries. New England Journal of Medicine, 352(15): 1514-1516.
6. Clem’s, M. L., Manıamfu, P., Louıson, D. K. Five Machine Learning Supervised Algorithms for The Analysis and the Prediction of Obesity. International Journal of Innovative Science and Research Technology, 7(12):1956-1964.
7. Cui, T., Chen, Y., Wang, J., Deng, H., Huang, Y. (2021, May). Estimation of Obesity levels based on Decision trees. In 2021 International Symposium on Artificial Intelli-gence and its Application on Media (ISAIAM) (pp. 160-165). IEEE.
8. Danacı, Ç., Avcı, D. Arslan Tuncer, S. (2023). Komşuluk Bileşen Analizi Tabanlı Makine Öğrenimi Yöntemleri ile Obezite Seviyelerinin Tahmini. Fırat Üniversitesi Mühendislik Bilimleri Dergisi, 35 (2): 433-442.

9. Dugan, T. M., Mukhopadhyay, S., Carroll, A., Downs, S. (2015). Machine learning techniques for prediction of ear-ly childhood obesity. Applied clinical informatics, 6(3):506-520.
10. Feller, W. (1991). An introduction to probability theory and its applications, John Wiley & Sons.Ltd, England. Ferdowsy, F., Rahi, K. S. A., Jabiullah, M. I., Habib, M. T. (2021). A machine learning approach for obesity risk pre-diction. Current Research in Behavioral Sciences, 2: 100053; DOI 10.1016/j.crbeha.2021.100053.
11. Jeon, J., Lee, S., Oh, C. (2023). Age-specific risk factors for the prediction of obesity using a machine learning ap-proach. Frontiers in Public Health, 10; DOI 10.3389/fpubh.2022.998782.
12. Koo, H. C., Tan, L. K., Lim, G. P., Kee, C. C., Omar, M. A. (2023). Obesity and Its Association with Undiagnosed Di-abetes Mellitus, High Blood Pressure and Hypercholes-terolemia in the Malaysian Adult Population: A National Cross-Sectional Study Using NHMS Data. International Journal of Environmental Research and Public Health, 20(4); DOI 10.3390/ijerph20043058.
13. Lever, J. (2016). Classification evaluation: It is important to understand both what a classification metric expresses and what it hides. Nature methods, 13(8): 603-605.
14. Mingers, J. (1989). An empirical comparison of pruning methods for decision tree induction. Machine learning, 4: 227-243.
15. Mondal, P. K., Foysal, K. H., Norman, B. A., Gittner, L. S. (2023). Predicting Childhood Obesity Based on Single and Multiple Well-Child Visit Data Using Machine Learn-ing Classifiers. Sensors, 23(2): 759; DOI 10.3390/s23020759.
16. Montañez, C. A. C., Fergus, P., Hussain, A., Al-Jumeily, D., Abdulaimma, B., Hind, J., Radi, N. (2017). Machine learn-ing approaches for the prediction of obesity using publicly available genetic profiles. 2017 International Joint Confer-ence on Neural Networks (IJCNN), May 14-19,2017, Online,IEEE, 2743-2750.
17. Moosavian, A., Ahmadi, H., Tabatabaeefar, A., Khazaee, M. (2013). Comparison of two classifiers; K-nearest neighbor and artificial neural network, for fault diagnosis on a main engine journal-bearing. Shock and Vibration, 20(2), 263-272.
18. Ogden, C. L., Carroll, M. D., Curtin, L. R., McDowell, M. A., Tabak, C. J., Flegal, K. M. (2006). Prevalence of over-weight and obesity in the United States, 1999-2004.
19. Pal, M., Mather, P. M. (2003). An assessment of the effec-tiveness of decision tree methods for land cover classifi-cation. Remote sensing of environment, 86(4): 554-565.
20. Quinlan, J. R. (2014). C4. 5: programs for machine learning. Elsevier.
21. Roy, M., Das, S., & Protity, A. T. (2023). OBESEYE: Inter-pretable Diet Recommender for Obesity Management us-ing Machine Learning and Explainable AI. arXiv preprint. 4(6); DOI https://doi.org/10.48550/arXiv.2308.02796.
22. Singh, B., Tawfik, H. (2020). Machine learning approach for the early prediction of the risk of overweight and obesity in young people. In Computational Science–ICCS 2020: 20th International Conference, June 3–5, 2020, Amster-dam, The Netherlands, 523-535.
23. Susmaga, R. (2004). Confusion matrix visualization. In Intelligent Information Processing and Web Mining, May 17–20,2004, Berlin, Heidelberg.107-1116.
24. UCI (2023). https://archive.ics.uci.edu/ (Erişim Tarihi: 10.06.2023) URL-1 (2023). https://www.researchgate.net/publication/278050782_The_use_of_the_k_nearest_neighbor_method_to_classify_the_representative_elements
25. URL-2 (2023). https://pyimagesearch.com/2021/05/24/grid-search-hyperparameter-tuning-with-scikit-learn-gridsearchcv/
26. Vizmanos, B., Cascales, A. I., Rodríguez‐Martín, M., Salme-rón, D., Morales, E., Aragón‐Alonso, A., Garaulet, M. (2023). Lifestyle mediators of associations among siestas, obesity, and metabolic health. Obesity, 31(5): 1227-1239.
27. Wanjau, M. N., Kivuti-Bitok, L. W., Aminde, L. N., & Veer-man, J. L. (2023). The health and economic impact and cost effectiveness of interventions for the prevention and control of overweight and obesity in Kenya: a stakeholder engaged modelling study. Cost Effectiveness and Re-source Allocation, 21(1): 1-21.
28. WHO (2022). European regional obesity report 2022. https://www.who.int/europe/publications/i/item/9789289057738 (Erişim Tarihi: 15.07.2023)
29. Yagin FH, Gülü M, Gormez Y, Castañeda-Babarro A, Colak C, Greco G, Fischetti F, Cataldi S.(2023) Estimation of Obesity Levels with a Trained Neural Network Approach optimized by the Bayesian Technique. Applied Sciences. 13(6); DOI 10.3390/app13063875
30. Young, S. R., Rose, D. C., Karnowski, T. P., Lim, S. H., Pat-ton, R. M. (2015). Optimizing deep learning hyper-parameters through an evolutionary algorithm. MLHPC '15: Proceedings of the Workshop on Machine Learning in High-Performance Computing Environments, Nov 15,2015, New York,United States,1-15.
31. Zheng, Z., Ruggiero, K. (2017). Using machine learning to predict obesity in high school students. In 2017 IEEE In-ternational Conference on Bioinformatics and Biomedi-cine (BIBM), Nov 13-16,2017, Online,IEEE. 2132-2138.