Scholarlink

<p>Assessing poverty or disparity requires accurate data that is geographically pertinent, trackable (real-time evaluation), and less error-prone, which is important for achieving the Sustainable Development Goals (SDG) of the United Nations to remove poverty. Acquiring these datasets is tough, particularly in developing nations with fluctuating economic dynamics. Although poverty is a complicated phenomenon it cannot be conceptually or realistically quantified by a single data category, therefore an attempt has been made using multisource information to quantify poverty or disparity. The present study aims at the integration of conventional and non-conventional datasets by using Geographic Information Systems (GIS) analysis and a machine learning-based Random Forest Regression (RFR) model to predict inequality for the Grid scale (10 × 10&#xa0;km) at a sub-national level. The significance of this research could substantially impact poverty alleviation efforts, which can provide valuable insight into wealth inequality that can guide evidence-based policy decisions, optimize resource allocation strategies, and achieve SDGs. The Demographic and Health Surveys (DHS), household wealth index (WI) factor score was considered to predict wealth inequality. To evaluate the model's efficacy in forecasting the WI factor score for India, the Random Forest Regression (RFR) model was supplied with both geospatial-based socio-economic datasets. The analysis revealed an R-square value of 0.86 between the observed and projected WI factor scores, demonstrating the model's high precision in predicting the WI. Besides, this study also discovered a negative association with a correlation (-0.6) between the district average WI and the Multidimensional Poverty Index (MPI) of India. Gini significance analysis determined crucial factors that are causing wealth disparity. Population count and POI density were shown to be the two most important variables, explaining 36.09% and 26.1% of the explanatory power, respectively. This suggests that places with a higher population density and POI density have more wealth disparity. Overall, the findings give essential insights into the factors that drive India's wealth inequality, which would help reform policies to eliminate this disparity.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Integrating Multi-Source Satellite Imagery and Socio-Economic Household Data for Wealth-Based Poverty Assessment of India: A GIS and Machine Learning Based Approach

  • 期刊论文

  • 发表日期：

    2025年6月2日

查看全文
  • Prashant Kumar Arya,
  • Koyel Sur,
  • Siddharth Dhote,
  • Harsh Siral,
  • Tanushree Kundu,
  • Balwant Singh Mehta,
  • Ravi Srivastava

摘要

Assessing poverty or disparity requires accurate data that is geographically pertinent, trackable (real-time evaluation), and less error-prone, which is important for achieving the Sustainable Development Goals (SDG) of the United Nations to remove poverty. Acquiring these datasets is tough, particularly in developing nations with fluctuating economic dynamics. Although poverty is a complicated phenomenon it cannot be conceptually or realistically quantified by a single data category, therefore an attempt has been made using multisource information to quantify poverty or disparity. The present study aims at the integration of conventional and non-conventional datasets by using Geographic Information Systems (GIS) analysis and a machine learning-based Random Forest Regression (RFR) model to predict inequality for the Grid scale (10 × 10 km) at a sub-national level. The significance of this research could substantially impact poverty alleviation efforts, which can provide valuable insight into wealth inequality that can guide evidence-based policy decisions, optimize resource allocation strategies, and achieve SDGs. The Demographic and Health Surveys (DHS), household wealth index (WI) factor score was considered to predict wealth inequality. To evaluate the model's efficacy in forecasting the WI factor score for India, the Random Forest Regression (RFR) model was supplied with both geospatial-based socio-economic datasets. The analysis revealed an R-square value of 0.86 between the observed and projected WI factor scores, demonstrating the model's high precision in predicting the WI. Besides, this study also discovered a negative association with a correlation (-0.6) between the district average WI and the Multidimensional Poverty Index (MPI) of India. Gini significance analysis determined crucial factors that are causing wealth disparity. Population count and POI density were shown to be the two most important variables, explaining 36.09% and 26.1% of the explanatory power, respectively. This suggests that places with a higher population density and POI density have more wealth disparity. Overall, the findings give essential insights into the factors that drive India's wealth inequality, which would help reform policies to eliminate this disparity.