A study on Hierarchical Clustering and the Distance metrics for Identifying Architectural Styles

With rapid advances in information technology, explosive growth is recorded in data generation and data collection capabilities in all areas. For this purpose, it is also necessary in an adjacent field such as constructing a solution to facilitate the detection of architectural styles of buildings, so developing a solution for rapid identification of architectural style would add value to smart city. Keeping in mind the numerous types of buildings and their appurtenance in images, the identification of style represents an important problem to solve. The developed application that aims at studying clustering efficiency is based on CBIR. This study reviews the most used distances (Euclidean and Manhattan) in the Data Mining techniques, providing an overview of the impact on the classification.