Regresyon ve Açıklanabilir Makine Öğrenmesi Kullanılarak Ülkeler Arası Atık Geri Dönüşüm Performansının Tahmin Edilmesi

Öz

Geri dönüşüm, sürdürülebilir atık yönetiminin ve döngüsel ekonominin temel unsurlarından biridir; ancak geri dönüşüm performansı ülkeler arasında önemli ölçüde farklılık göstermektedir. Bu çalışma, kentsel atıkların geri dönüşüm performansıyla ilişkili ekonomik ve atık kaynaklı faktörleri incelemekte ve makine öğrenmesi (ML) modellerinin geleneksel istatistiksel modellemeye kıyasla tahmin performansını iyileştirip iyileştirmediğini değerlendirmektedir. Analiz, Dünya Bankası What a Waste 3.0 verileri kullanılarak nüfusu en az bir milyon olan 84 ülkeyi kapsamaktadır. Kentsel atıkların geri dönüşüm oranı; kişi başına GSYİH, kişi başına kentsel katı atık (KKA) üretimi ile gıda, plastik ve kâğıt-karton atık oranları kullanılarak modellenmiştir. Çoklu doğrusal regresyon (MLR); Random Forest (RF), XGBoost ve Destek Vektör Regresyonu (SVR) ile iç içe tekrarlı beş katlı çapraz doğrulama kullanılarak karşılaştırılmış; en iyi performans gösteren ML modelinin yorumlanmasında SHapley Additive exPlanations (SHAP) yöntemi kullanılmıştır. Kişi başına GSYİH’nın doğal logaritması [ln(GSYİH per capita)], MLR modelindeki istatistiksel olarak anlamlı tek bağımsız yordayıcı olmuş ve küresel SHAP öneminin %47,67'sini oluşturarak RF modelinde de baskın yordayıcı olarak belirlenmiştir. Gıda atığı (%18,95), kâğıt-karton (%14,71) ve KKA üretimi (%14,69) de anlamlı düzeyde katkı sağlarken plastik atığın önemi sınırlı kalmıştır (%3,98). RF en güçlü tahmin performansını (R²=0,312) göstererek MLR'yi (R²=0,229) geride bırakmış; XGBoost sınırlı bir iyileşme (R²=0,263) sağlarken SVR (R²=0,215) MLR'den daha düşük performans göstermiştir. Gelir gruplarına göre gerçekleştirilen SHAP analizi, GSYİH'nın tüm gelir gruplarında baskın yordayıcı olmaya devam ettiğini, atıkla ilişkili değişkenlerin göreli öneminin ise farklılık gösterdiğini ortaya koymuştur. Bulgular, açıklanabilir makine öğrenmesinin geleneksel istatistiksel modellemeyi tamamlayıcı bir yaklaşım olarak değerini ortaya koymaktadır.

Anahtar Kelimeler

• Geri dönüşüm performansı
• Kentsel katı atıklar
• Makine öğrenmesi
• Random Forest
• SHAP

Predicting Cross-Country Waste Recycling Performance Using Regression and Explainable Machine Learning

Öz

Recycling is central to sustainable waste management and the circular economy, yet performance varies substantially across countries. This study examines economic and waste-related factors associated with municipal recycling performance and evaluates whether machine learning (ML) models improve prediction relative to conventional statistical modelling. The analysis included 84 countries with populations of at least one million using World Bank What a Waste 3.0 data. Municipal recycling rate was modelled using GDP per capita, municipal solid waste (MSW) generation per capita, and food, plastic, and paper and cardboard waste shares. Multiple linear regression (MLR) was compared with Random Forest (RF), XGBoost, and Support Vector Regression (SVR) using nested repeated five-fold cross-validation, while SHapley Additive exPlanations (SHAP) were used to interpret the best-performing ML model. ln(GDP per capita) was the only statistically significant independent predictor in MLR and the dominant RF predictor, accounting for 47.67% of global SHAP importance. Food waste (18.95%), paper and cardboard (14.71%), and MSW generation (14.69%) also contributed meaningfully, whereas plastic waste showed limited importance (3.98%). RF achieved the strongest predictive performance (R2=0.312), outperforming MLR (R2=0.229), while XGBoost showed a modest improvement (R2=0.263) and SVR did not (R2=0.215). Income-stratified SHAP analysis showed that GDP remained dominant across income groups, while waste-related importance varied. These findings demonstrate the value of explainable ML as a complement to conventional statistical modelling.

Anahtar Kelimeler

• Recycling performance
• Municipal solid waste
• Machine learning
• Random Forest
• SHAP

Kaynakça

1. Cehlár, M., Taušová, M., Ivanková, V., & Khouri, S. (2025). Municipal waste recycling in the EU: A multi-method analysis of determinants and country profiles (2005–2023). Frontiers in Environmental Science, 13, Article 1670365. https://doi.org/10.3389/fenvs.2025.1670365
2. Choudhury, M., Suhag, M., Jha, P., Harne, K. R., Samanta, P., Choudhury, M., Borgohain, R. B., Majumdar, S., Kravchenko, Y., & Choudhary, R. (2025). Urban waste management and challenges in emerging smart city systems governance and circular economy perspectives. Ecological Engineering & Environmental Technology, 26(10), 243–266. https://doi.org/10.12912/27197050/211014
3. Çevik, M. (2025). Akıllı kentsel atık yönetimi için dijital ikiz teknolojisi [Digital twin technology for smart urban waste management]. MCBÜ Soma Meslek Yüksekokulu Teknik Bilimler Dergisi, 2(40), 41–53. https://doi.org/10.47118/somatbd.1836379
4. Ellen MacArthur Foundation. (2019). Completing the picture: How the circular economy tackles climate change. https://www.ellenmacarthurfoundation.org/completing-the-picture
5. Hussein, N. A.-H. K., Huang, X., Li, X., & Yao, J. (2024). Assessing the impact of circular economy practices on global waste management systems. Estidamaa, 2024, 23–30.
6. Kostakis, I., Halkos, G., & Sardianou, E. (2026). Sustainability in municipal waste recycling: An empirical investigation of economic and institutional determinants. Global Economics Research, 2(1), Article 100018. https://doi.org/10.1016/j.ecores.2025.100018
7. Leclerc, S. H., & Badami, M. G. (2024). Extended producer responsibility: An empirical investigation into municipalities' contributions to and perspectives on e-waste management. Environmental Policy and Governance, 34(2), 111–124. https://doi.org/10.1002/eet.2059
8. lo Storto, C. (2025). Evaluating the eco-efficiency of municipal solid waste management: Determinants, paradoxes, and trade-offs. Urban Science, 9(10), Article 395. https://doi.org/10.3390/urbansci9100395

9. Nimjerdi, A. H. (2026). The impact of open innovation on municipal waste recycling rate and resource productivity in the European Union countries: The moderating role of environmental regulations. Sustainable Environment, 12(1), Article 2655489. https://doi.org/10.1080/27658511.2026.2655489
10. Osińska, M. (2024). The determinants of municipal solid waste management efficiency in EU countries. Economics and Environment, 88(1), 1–14. https://doi.org/10.34659/eis.2024.88.1.637
11. Rolewicz-Kalińska, A., Lelicińska-Serafin, K., & Manczarski, P. (2020). The circular economy and organic fraction of municipal solid waste recycling strategies. Energies, 13(17), Article 4366. https://doi.org/10.3390/en13174366
12. Rondón Toro, E., Lobo, A., & Gallardo Izquierdo, A. (2022). Circularity indicator for municipal solid waste treatment plants. Journal of Cleaner Production, 380, Article 134807. https://doi.org/10.1016/j.jclepro.2022.134807
13. Rutkowski, J. E. (2020). Inclusive packaging recycling systems: Improving sustainable waste management for a circular economy. Detritus, 13, 29–46. https://doi.org/10.31025/2611-4135/2020.14037
14. Tunalı, D. (2026). Assessing waste treatment efficiency of European Union countries. Environmental Research and Technology, 9(3), 513–519. https://doi.org/10.35208/ert.1569431
15. Tüzüner, Z., & Alp, İ. (2018). Comparison of solid waste management performances of Turkey and EU countries associated with Malmquist index. Journal of Polytechnic, 21(1), 75–81. https://doi.org/10.2339/politeknik.386857
16. United Nations Environment Programme. (2024). Global waste management outlook 2024. United Nations Environment Programme. https://www.unep.org/resources/global-waste-management-outlook-2024
17. Velis, C. A., Wilson, D. C., Gavish, Y., Grimes, S. M., & Whiteman, A. (2023). Socio-economic development drives solid waste management performance in cities: A global analysis using machine learning. Science of the Total Environment, 872, Article 161913. https://doi.org/10.1016/j.scitotenv.2023.161913
18. World Bank. (2026). What a Waste 3.0: Global snapshot of solid waste management toward circularity until 2050. World Bank. https://doi.org/10.1596/978-1-4648-2309-1
19. Ye, M., Jin, Y., & Deng, F. (2022). Municipal waste treatment efficiency in 29 OECD countries using three-stage Bootstrap-DEA model. Environment, Development and Sustainability, 24(9), 11369–11391. https://doi.org/10.1007/s10668-022-02227-4
20. Yüce Gök, N. C., & Güreşçi, G. (2025). Circular economy and waste management: Effects of the waste management on economic growth in the European Union and Turkiye. Beykoz Akademi Dergisi, 13(2), 586–604. https://doi.org/10.14514/beykozad.1630486
21. Zon, J. L. N., Leopoldino, C. J., Yamane, L. H., & Siman, R. R. (2020). Waste pickers organizations and municipal selective waste collection: Sustainability indicators. Waste Management, 118, 219–231. https://doi.org/10.1016/j.wasman.2020.08.023