Impediments to Machine Learning Adoption for Land Use Change Prediction: Perceptions of Registered Land Use Professionals in Akure, Nigeria

Abstract

Recent advances in machine learning (ML) have created new opportunities for improving the prediction of urban land use change. Despite these prospects, the adoption of ML techniques among land use professionals in Nigerian cities remains limited due to several structural, institutional, technical, and resource-related challenges. This study examined the barriers to the adoption of machine learning for land use change prediction in Akure, Ondo State, Nigeria, from the perspective of registered land use professionals. A census survey was conducted among 86 professionals comprising Estate Surveyors and Valuers, Town Planning officials, Land Surveyors, and GIS/Remote Sensing experts, out of which 64 valid responses were retrieved, representing a 74% response rate. Data obtained from the survey were analyzed using descriptive statistics, including frequency distribution, percentages, Relative Importance Index (RII), and Weighted Mean Score (WMS). Twenty-two identified barriers were grouped into four categories: data-related barriers, model-related barriers, policy and institutional barriers, and computational/other barriers. The findings from the study revealed a Grand RII of 0.8338, indicating that respondents generally perceived the identified barriers as highly significant. Poor internet connectivity (RII = 0.9250) and funding/resource constraints (RII = 0.9250) ranked as the most severe impediments to ML adoption, followed by inadequate computing infrastructure (RII = 0.9031) and continued dependence on traditional methods (RII = 0.8938). The study concludes that infrastructural and resource deficiencies, rather than professional resistance are the major factors limiting the adoption of ML tools in the study area. The research therefore recommended improved digital infrastructure, increased institutional support, enhanced professional training, and better access to geospatial data to facilitate the integration of ML into land use planning practice in Nigeria.