Scholarlink

<p>The urban growth prediction is essential for sustainable urban planning in rapidly urbanizing cities like Lucknow, India. In this study, three machine learning models, Support Vector Machine (SVM), Random Forest (RF), and Artificial Neural Network (ANN) are evaluated in predicting the urban growth patterns from 2021 to 2031 using geospatial data comprising environmental and socioeconomic variables. The models generated urban growth probability maps that classified the study area into five probability classes very low, low, medium, high and very high. The RF model showed the highest accuracy (87.69%) and precision (75.36% for urban areas), and therefore proved to be the most appropriate model for localized and stable urban growth prediction. While the SVM model was effective at detecting emerging urbanized areas, it had a strong recall (74.58%), but at the cost of precision. The ANN model had the highest recall (78.25%) and the best ability to identify dispersed growth patterns in peri urban zones. This work highlights the application of machine learning in urban growth modeling and provides scalable methods for other urbanizing regions. The results offer essential lessons for data driven decision making which help in achieving sustainable urban development, balancing growth with environment and social factors.</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

Decoding urban expansion: a machine learning perspective on Lucknow's growth trajectory

  • 期刊论文

  • 发表日期：

    2025年3月10日

查看全文
  • Danish Khan,
  • Nizamuddin Khan

摘要

The urban growth prediction is essential for sustainable urban planning in rapidly urbanizing cities like Lucknow, India. In this study, three machine learning models, Support Vector Machine (SVM), Random Forest (RF), and Artificial Neural Network (ANN) are evaluated in predicting the urban growth patterns from 2021 to 2031 using geospatial data comprising environmental and socioeconomic variables. The models generated urban growth probability maps that classified the study area into five probability classes very low, low, medium, high and very high. The RF model showed the highest accuracy (87.69%) and precision (75.36% for urban areas), and therefore proved to be the most appropriate model for localized and stable urban growth prediction. While the SVM model was effective at detecting emerging urbanized areas, it had a strong recall (74.58%), but at the cost of precision. The ANN model had the highest recall (78.25%) and the best ability to identify dispersed growth patterns in peri urban zones. This work highlights the application of machine learning in urban growth modeling and provides scalable methods for other urbanizing regions. The results offer essential lessons for data driven decision making which help in achieving sustainable urban development, balancing growth with environment and social factors.