A Machine-Insight-Based Geocomputational Approach to Decode the Complexities of the Restless Landscapes: Toward Integration of Imaginations and Planning Interventions

The proliferating urban influence brought a complex sociospatial, economic structure over the urban periphery. Previous studies focused on different approaches to capturing this intricate process; generalization was done by setting their own criteria. This study builds a decision-making model to comprehend the nature of Varanasi city’s periphery. Settlement-wise data are collected to calculate the study’s initial parameters: functional index, physical infrastructure dynamics score, physiological density, and nonagricultural land to total land ratio. The independent component analysis calculates the composite score. The learning cloud incorporates settlement-wise data in an encoded format to run the XGBoost machine learning algorithm. The findings highlight a competitive pattern of settlements and identify four periurban zones based on the deviation diffusion of calculated parameters. The study implements the Monte Carlo simulation for visualizing periurban decision space and recorded an asymptotic one-sample Kolmogorov–Smirnov test score of 0.051 at a p value of 0.076. The findings suggest four peripheral settlement clusters where the Gaussian mixture model confirms the reliability. This study highlights the importance of these clusters in attracting government and private players to invest and innovate functioning urban like institutions. This study’s methodological framework and inclusive approaches for city region planning and policy formulation remain significant.

不断扩张的城市影响力在城市边缘区催生了复杂的社会空间与经济结构。既往研究围绕捕捉这一复杂过程的不同路径展开，均通过设定各自的判定标准实现研究泛化。本研究构建决策模型以解析瓦拉纳西（Varanasi）城市边缘区的本质特征。研究按聚落收集数据，用于计算初始参数：功能指数、实体基础设施动态得分、生理密度，以及非农业用地占总用地比例。采用独立成分分析（Independent Component Analysis）计算综合得分。学习云平台以编码格式整合聚落数据，以运行XGBoost机器学习算法。研究结果揭示了聚落的竞争性分布格局，并基于计算参数的偏差扩散特征识别出4个城市边缘区。本研究采用蒙特卡洛（Monte Carlo）模拟对城市边缘区决策空间进行可视化，所得渐近单样本柯尔莫哥洛夫-斯米尔诺夫（Kolmogorov–Smirnov）检验得分为0.051，对应p值为0.076。研究结果划定了4个边缘聚落集群，高斯混合模型（Gaussian Mixture Model）验证了该集群划分的可靠性。本研究强调了此类集群在吸引政府与市场主体投资、构建类城市运行机制并推动其创新方面的重要价值。本研究的方法论框架与面向城市区域规划、政策制定的包容性研究路径，仍具有重要参考意义。