A Data Mining Approach to Uncovering Hidden Patterns in Islamic Architecture

Given the growing significance of large-scale architectural databases and digital heritage archives, the discovery of meaningful patterns and the application of intelligent analysis methods have become increasingly relevant in the study of architectural history. Traditional data mining techniques have primarily focused on identifying associations among items within transactional data. Among these, the extraction of fuzzy association rules has emerged as a promising approach for handling imprecise or overlapping information, which is often the case in architectural datasets. This paper introduces a novel method, RM (Rapid Miner), designed to efficiently discover fuzzy association rules by categorizing transaction records based on their lengths and scanning the database only once. Each record is assigned to a specific table depending on its size, and large itemsets are then generated from these categorized tables. When applied to the study of Islamic architecture, this approach allows for the identification of design patterns and stylistic elements that may not be strictly defined, such as variations in geometric motifs, spatial organization, or decorative elements. The method proves effective in revealing subtle relationships and trends across different periods and regions, offering valuable insights into the evolution and diversity of Islamic architectural heritage. Experimental results confirm the improved speed and efficiency of the proposed method compared to conventional algorithms.

کلیدواژه ها:

Association Rules ، Fuzzy ، Fuzzy Itemsets