The Reflection of Traditional Architectural Features in Modern Residence using Machine Learning: A case study near ancient town in Shanghai

This study examines the role of the machine learning model in the clustering of groups having similarities in rating architectural styles. The survey approach was used to collect data from participants residing near Qibao Ancient Town in Shanghai, China. The collected data was preprocessed for extraction of significant architectural features. The extracted information was used to train and evaluate the K-Means model to cluster residents' satisfaction and identify their needs. Results show that moderate architectural similarity (0.4 - 0.8) was achieved by the machine learning model, while human-rated similarity (0.0 - 1.0) was obtained in this study. The K-Means clustering approach categorized participants into three clusters based on their satisfaction and their concerns on lighting and noise. This research study suggests redesigning fixtures with tactile finishes, ergonomic handles, and modular shelving. Furthermore, this work suggests installing LED lighting in cold regions and adding acoustic panels in high-traffic areas.