한국지리정보학회지 (Journal of the Korean Association of Geographic Information Studies)

  • 제26권4호
  • /
  • Pages.97-115
  • /
  • 2023
  • /
  • 1226-9719(pISSN)
  • /
  • 2287-6952(eISSN)
한국지리정보학회 (The Korean Association of Geographic Information Studies)
권호 표지
DOI QR코드

DOI QR Code

해석가능한 기계학습을 적용한 소지역 인구 추정에 관한 연구: 부산광역시를 대상으로

A Study on the Population Estimation of Small Areas using Explainable Machine Learning: Focused on the Busan Metropolitan City

  • 김유현 (부산대학교 공과대학 도시공학과) ;
  • 김동현 (부산대학교 공과대학 도시공학과)
  • Yu-Hyun KIM (Department of Urban Planning and Engineering, Pusan National University) ;
  • Donghyun KIM (Department of Urban Planning and Engineering, Pusan National University)
  • 투고 : 2023.11.02
  • 심사 : 2023.11.16
  • 발행 : 2023.12.31
PDF 다운로드
⟨ 이전 논문 다음 논문 ⟩

초록

최근 저출산, 고령화 등 인구의 구조가 급격히 변화하고 있고 인구 분포의 불균등성이 확대되고 있는 시점에서 인구 추정 방식의 변화가 요구되고 있으며 소지역 단위에서 보다 정확한 추정이 요구되고 있다. 본 연구는 이러한 인구 추정 방식 변화 요구에 대응하기 위해 부산광역시를 대상으로 해석가능한 기계학습 방법을 적용하여 500m 격자 단위에서 2040년 인구를 추정하는 것을 목적으로 하고 있다. 해석가능한 기계학습의 방법과 코호트 요인법을 각각 적용하여 격자별 인구추정 결과를 비교해본 결과, 기계학습 방법이 인구 구조 변동에 영향을 미칠 가능성이 있는 여러 변수의 조합 반영이 가능하여 보다 낮은 오차를 도출함으로써 소지역과 같이 인구 변화폭이 큰 지역의 추정에 있어 적용력이 높음을 확인하였다. 인구감소시대에 과대추정된 인구 값은 도시계획에서 투자의 비효율성과 특정 부문에 대한 과잉 투자에 따른 타 부문에서의 질적 저하와 같은 문제를 일으킬 가능성이 높으며, 과소추정된 인구 값 역시 도시의 축소를 가속화시켜 삶의 질을 저하시키는 문제를 초래하므로 적절한 인구 추정 방법과 대안을 마련해야 할 필요가 있을 것으로 판단된다.

In recent years, the structure of the population has been changing rapidly, with a declining birthrate and aging population, and the inequality of population distribution is expanding. At this point, changes in population estimation methods are required, and more accurate estimates are needed at the subregional level. This study aims to estimate the population in 2040 at the 500m grid level by applying an explainable machine learning to Busan in order to respond to this need for a change in population estimation method. Comparing the results of population estimation by applying the explainable machine learning and the cohort component method, we found that the machine learning produces lower errors and is more applicable to estimating areas with large population changes. This is because machine learning can account for a combination of variables that are likely to affect demographic change. Overestimated population values in a declining population period are likely to cause problems in urban planning, such as inefficiency of investment and overinvestment in certain sectors, resulting in a decrease in quality in other sectors. Underestimated population values can also accelerate the shrinkage of cities and reduce the quality of life, so there is a need to develop appropriate population estimation methods and alternatives.

키워드

과제정보

이 논문은 2023년 대한민국 교육부와 한국연구재단의 지원(NRF-2020S1A3A2A01095064)과 국토교통부의 스마트시티 혁신인재육성사업으로 지원을 받아 수행되었습니다.

참고문헌

  1. Baker, J., A. Alcantara., X. Ruan., K. Watkins., and S. Vasan. 2013. A comparative evaluation of error and bias in census tract-level age/sex-specific population estimates: component I (net-migration) vs component III (Hamilton-Perry). Population Research and Policy Review 32:919-942. https://doi.org/10.1007/s11113-013-9295-4
  2. Baker, J., A. Alcantara., X. Ruan., K. Watkins., and S. Vasan, 2014. Spatial weighting improves accuracy in small-area demographic forecasts of urban census tract populations. Journal of Population Research 31:345-359. https://doi.org/10.1007/s12546-014-9137-1
  3. Baker, J., D. Swanson., J. and Tayman. 2021. The accuracy of Hamilton-Perry population projections for census tracts in the United States. Population Research and Policy Review 40:1341-1354. https://doi.org/10.1007/s11113-020-09601-y
  4. Breidenbach, P., M. Kaeding., and S. Schaffner. 2019. Population projection for Germany 2015-2050 on grid level (RWI-GEO-GRID -POP-Forecast). Jahrbucher fur Nationalokonomie und Statistik 239(4):733-745. https://doi.org/10.1515/jbnst-2017-0149
  5. Cai, Q. 2007. New techniques in small area population estimates by demographic characteristics. Population Research and Policy Review 26:203-218. https://doi.org/10.1007/s11113-007-9028-7
  6. Caraviello, D. Z., K. A. Weigel., M. Craven., D. Gianola., N. B. Cook., K. V. Nordlund,, ... and M. C. Wiltbank. 2006. Analysis of reproductive performance of lactating cows on large dairy farms using machine learning algorithms. Journal of dairy science 89(12):4703-4722. https://doi.org/10.3168/jds.S0022-0302(06)72521-8
  7. Chen, T., and C. Guestrin. 2016. Xgboost: A scalable tree boosting system. In Proceedings of the 22nd acm sigkdd international conference on knowledge discovery and data mining. 785-794.
  8. Chen, Y., F. Guo., J. Wang., W. Cai., C. Wang., K. and Wang. 2020. Provincial and gridded population projection for China under shared socioeconomic pathways from 2010 to 2100. Scientific Data 7(1):83.
  9. Chen, Y., X. Li., K. Huang., M. Luo., and M. Gao. 2020. High-resolution gridded population projections for China under the shared socioeconomic pathways. Earth's Future 8(6):e2020EF001491.
  10. Chi, G., and P. R. Voss. 2011. Small-area population forecasting: Borrowing strength across space and time. Population, Space and Place 17(5):505-520. https://doi.org/10.1002/psp.617
  11. Davis, H. C. 1995. Demographic projection techniques for regions and smaller areas: a primer. UBC Press.
  12. Ford, A., S. Barr., R. Dawson., J. Virgo., M. Batty., and J. Hall. 2019. A multi-scale urban integrated assessment framework for climate change studies: A flooding application. Computers, Environment and Urban Systems 75:229-243. https://doi.org/10.1016/j.compenvurbsys.2019.02.005
  13. Goodchild, M. F., L. Anselin., and U. Deichmann. 1993. A framework for the areal interpolation of socioeconomic data. Environment and planning A. 25(3):383-397. https://doi.org/10.1068/a250383
  14. Lundberg, P., and P. Frodin. 1998. Ecosystem resilience and productivity: are predictions possible?. Oikos 603-606.
  15. McKee, J. J., A. N. Rose., E. A. Bright., T. Huynh., and B. L. Bhaduri. 2015. Locally adaptive, spatially explicit projection of US population for 2030 and 2050. Proceedings of the National Academy of Sciences 112(5):1344-1349. https://doi.org/10.1073/pnas.1405713112
  16. Miller, T. 2019. Explanation in artificial intelligence: Insights from the social sciences. Artificial intelligence 267:1-38. https://doi.org/10.1016/j.artint.2018.07.007
  17. Perry, C. 2013. The neighborhood unit. In The urban design reader. 98-109. Routledge.
  18. Rayer, S. 2007. Population forecast accuracy: does the choice of summary measure of error matter?. Population Research and Policy Review 26:163-184. https://doi.org/10.1007/s11113-007-9030-0
  19. Samuel, A. L. 1959. Some studies in machine learning using the game of checkers. IBM Journal of research and development 3(3):210-229. https://doi.org/10.1147/rd.33.0210
  20. Shapley, L. S. 1997. A value for n-person games. Classics in game theory 69.
  21. Smith, S. K., and P. A. Morrison. 2005. Small-area and business demography. 761-785. Springer US.
  22. Smith, S. K., J. Tayman., and D. A. Swanson. 2005. State and local population projections: Methodology and analysis.
  23. Smith, S. K., J. Tayman., and D. A. Swanson. 2013. A practitioner's guide to state and local population projections. Springer Netherlands.
  24. Stillwell, J., and M. Clarke. (Eds.). 2011. Population dynamics and projection methods (Vol. 4). Springer Science & Business Media.
  25. Swanson, D. A. (Ed.). 2017. The frontiers of applied demography. Springer International Publishing.
  26. Swanson, D. A., A. Schlottmann., and B. Schmidt. 2010. Forecasting the population of census tracts by age and sex: An example of the Hamilton-Perry method in action. Population Research and Policy Review 29:47-63. https://doi.org/10.1007/s11113-009-9144-7
  27. Swanson, D. A., J. Tayman., and C. F. Barr. 2000. A note on the measurement of accuracy for subnational demographic estimates. Demography 37(2):193-201. https://doi.org/10.2307/2648121
  28. Triantakonstantis, D., and G. Mountrakis. 2012. Urban growth prediction: a review of computational models and human perceptions.
  29. Van Lent, M., W. Fisher., and M. Mancuso. 2004. An explainable artificial intelligence system for small-unit tactical behavior. In Proceedings of the national conference on artificial intelligence. 900-907. Menlo Park, CA; Cambridge, MA; London; AAAI Press; MIT Press; 1999.
  30. Weber, H. 2020. How well can the migration component of regional population change be predicted? A machine learning approach applied to German municipalities. Comparative Population Studies-Zeitschrift fur Bevolkerungswissenschaft. 45:143-178. https://doi.org/10.12765/CPoS-2020-08
  31. Wilson, T. 2015. New evaluations of simple models for small area population forecasts. Population. Space and Place 21(4):335-353. https://doi.org/10.1002/psp.1847
  32. Wilson, T., I. Grossman., and J. Temple. 2021. Evaluation of the best M4 competition methods for small area population forecasting. International Journal of Forecasting.
  33. Yan, X., X. Liu., and X. Zhao. 2020. Using machine learning for direct demand modeling of ridesourcing services in Chicago. Journal of Transport Geography 83:102661.
  34. Ahn, J.H. 2020. Explainable AI, Dissecting AI.
  35. Byeon, S.Y., D.C. Lee., and K.H. Kim. 2023. Estimating Gridded Population using Day-Time Satellite Image by CNN. Journal of The Korean Data Analysis Society 25(2):467-479. https://doi.org/10.37727/jkdas.2023.25.2.467
  36. Cho, D.H., and S.I. Lee. 2011. Population Projections for Busan Using a Biregional Cohort-Component Method. Journal of the Korean Geographical Society 46(2): 212-232.
  37. Choi, H.J., S.H. Choi., and S.J. Hong. 2019. Development of Estimation Method for Future Population by Eup-Myeon-Dong Unit. Journal of Real Estate Analysis 5(3):67-87. https://doi.org/10.30902/jrea.2019.5.3.67
  38. Kim, B.S., H.C. Jeon., and D.B. Shin. 2022. Development and Accuracy Analysis of the Population Estimation Models based on Building Volume and Landuse. Journal of Korean Society for Geospatial Information Science 30(4):25-34. https://doi.org/10.7319/kogsis.2022.30.4.025
  39. Kim, S.H. 2022. Empirical studies on the stock price prediction performance of Xgboost models. Journal of Social Science. 39(1):29-55.
  40. Lee, B.K. 2019. 2040 Future Population Projections Study. Korea Research Institute for Human Settlements.
  41. Lee, C.H., and S.I. Lee. 2006. Generation of Synthetic Population for Buildings. Journal of Korea Planning Association 41(6):37-50.
  42. Lee, J.J., Y.R. Lee., D.H. Lim., and H.C. Ahn. 2021. A Study on the Employee Turnover Prediction using XGBoost and SHAP. The Journal of Information Systems 30(4):21-42. https://doi.org/10.5859/KAIS.2021.30.4.21
  43. Lee, S.I., and D.H. Cho. 2020. Trend Extrapolation Methods for Small Area Population Projections in Korea. Journal of Geography Education 64:1-19.
  44. Lee, S.S. 2017. Demographic Dynamics and Policy Response. Health and Welfare Policy Forum 2017(1):29-46.
  45. Park, J.S., S.J. Kim., and S.G. Lee. 2022. Analysis of Determining Factors of Urban Vitality with Mobile Phone Location-Based Origin-Destination Bigdata by Travel Purpose : Using the PageRank Algorithm and SHAP Machine Learning. Journal of Korea Planning Association 57(5):72-89. https://doi.org/10.17208/jkpa.2022.10.57.5.72
  46. Woo, H.B., J.Y. Yang., S.H. Cho., and H.S. Ahn. 2016. Demographic forecasting: analytic review and assessment.
  47. Yang, G.P., and H.J. Chun. 2022. Analysis of the Relationship between Housing Prices and Regional Characteristics Using Machine Learning and XAI. Korea Real Estate Review 32(3):7-24. https://doi.org/10.35136/krer.32.3.1