違章建築對城市地區的公共安全、衛生和安全等問題常引起高度的關注。這些未經申請核准的建築物常常存在著事故或災害的風險。無論是違章建築頻繁發生的火災事件，亦或是違章工廠導致自然環境污染，這些問題對於市民大眾和政府都變得非常重要。本次研究為使用自組織映射群集演算法（SOMCM）針對臺北市自1995年以來的所有違章建築之數據進行分析，在與8名專家進行訪談後，確定了影響違章建築之處理方式的因素，包括行政區、違建位置、違建面積、違建材料和施工進度等變量。利用SOMCM演算法，研究發現：1.大多數違章建築屬於新違建、2.應該針對每個行政區域采用不同的處理策略、3.屋頂平台之違建是各違建位置中所占比例最高；本研究可讓政府更有信心將資源優先用於解決屋頂、陽台等位置之違建，將有限的資源運用在最需要的地方，減少不必要的資源浪費。;Illegal constructions have given rise to concerns regarding public safety, hygiene, and security in urban areas. There is a constant risk of accidents or hazards associated with these unauthorized structures. Whether it is frequent fire incidents in illegal buildings or the contamination of the natural environment caused by illegal factories, these issues have become a matter of significant importance for both the general public and the government. The research goal is aimed at providing efficient strategy using Self-Organizing Map-based Cluster Merging (SOMCM) algorithm. The data collection targets all illegal construction in Taipei City since 1995 when the regulations and database were renewed up-to-date. After interviewing 8 experts who determined impact factors for how to handle illegal construction, 6 variables are verified including Administrative District, Location, Construction Area, Construction Materials, and Construction Progress/Status. Utilizing the SOMCM algorithm, the study has found that: (1) The largest number of unauthorized constructions are of new origin; (2) Different strategies should be employed for the handling of unauthorized constructions in each administrative district; (3) Roof terraces have the highest proportion among unauthorized construction types. This can give the government more confidence in prioritizing resources for addressing unauthorized construction on roof terraces, for example, applying limited resources where they are most needed, and reducing unnecessary resource waste.
