Scholarlink

<p>Artificial Neural Networks (ANNs) have emerged as powerful tools for addressing complex socioeconomic challenges. In this context, our study explores the application of ANNs to model poverty, considering the unique characteristics and uncertainties associated with poverty assessment. Focusing on the Odisha state in India, we present an innovative approach that employs ANNs to predict poverty levels within intervals, acknowledging the inherent data uncertainties. We have prepared a custom cost function in the neural network that includes a trainable parameter <InlineEquation ID="IEq1"> <EquationSource Format="TEX">\(\delta \)</EquationSource> </InlineEquation> to determine the optimal interval width during backpropagation. The ANN architecture adopted in this study yields promising results. For the considered problem, the prediction interval width (<InlineEquation ID="IEq2"> <EquationSource Format="TEX">\(\delta = 1.29\)</EquationSource> </InlineEquation>) obtained indicates the model’s effectiveness. Also, the training and test losses were 0.0113 and 0.0284, respectively. By leveraging this adaptable ANN model, we can predict intervals for the headcount ratio (HR) across diverse regions within Odisha. Our research advances the synergy between advanced machine learning techniques and real-world social challenges, contributing to the ongoing efforts to combat poverty in Odisha and beyond.</p> Graphical abstract <p></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

Application of ANNs for predicting poverty levels with interval estimation: a case study of Odisha, India

  • 期刊论文

  • 发表日期：

    2026年1月8日

查看全文
  • Sandeep Kumar,
  • S. Chakraverty,
  • Narayan Sethi

摘要

Artificial Neural Networks (ANNs) have emerged as powerful tools for addressing complex socioeconomic challenges. In this context, our study explores the application of ANNs to model poverty, considering the unique characteristics and uncertainties associated with poverty assessment. Focusing on the Odisha state in India, we present an innovative approach that employs ANNs to predict poverty levels within intervals, acknowledging the inherent data uncertainties. We have prepared a custom cost function in the neural network that includes a trainable parameter \(\delta \) to determine the optimal interval width during backpropagation. The ANN architecture adopted in this study yields promising results. For the considered problem, the prediction interval width ( \(\delta = 1.29\) ) obtained indicates the model’s effectiveness. Also, the training and test losses were 0.0113 and 0.0284, respectively. By leveraging this adaptable ANN model, we can predict intervals for the headcount ratio (HR) across diverse regions within Odisha. Our research advances the synergy between advanced machine learning techniques and real-world social challenges, contributing to the ongoing efforts to combat poverty in Odisha and beyond.

Graphical abstract