Assessing smart cities' effectiveness: machine learning approaches

Amid the global emergence of smart cities, there exists a lack of consensus among scholars and city leaders regarding their perception and development. Notably, there is a dearth of quality indicators for evaluating the progress of smart city development. This study addresses this gap by focusing on identifying the drivers that influence residents' assessments of life quality and comfort. By gathering assessments from residents in priority areas identified as problematic for city prosperity, and incorporating basic measures of technological development, machine-learning models were constructed using RapidMiner Studio. These models aim to predict the Human Development Index (HDI) of the city and discern the most impactful drivers related to citizens' life satisfaction. The research compares various models, ultimately selecting the optimal Fast Large Margin model. The findings highlight crucial concerns for residents, including air pollution, recycling, basic amenities, and health services. The study relies on a unique dataset comprising official statistical information from 141 smart cities across 73 countries. The developed models offer valuable insights for decision-makers, enabling the formulation of effective strategies for sustainable smart city development and the enhancement of digitalization policies.