Development of the Small Space Analysis Method Considering Spatial Ambiguity Based on Semi-Supervised Clustering

2015 Fiscal Year Final Research Report

• Project/Area Number: 25420633
• Research Category: Grant-in-Aid for Scientific Research (C)
• Allocation Type: Multi-year Fund
• Section: 一般
• Research Field: Town planning/Architectural planning
• Research Institution: Osaka City University
• Principal Investigator: Takizawa Atsushi 大阪市立大学, 大学院工学研究科, 准教授 (40304133)
• Project Period (FY): 2013-04-01 – 2016-03-31
• Keywords: 空間分析 / イベント / 領域分割 / 半教師ありクラスタリング / 分類問題 / 混合整数2次計画問題

Outline of Final Research Achievements

In this study, I have developed a novel and general spatial analysis method for relatively small space such as architectures and streets. This analysis method can statistically explain and predict the place where human’s unconscious activities might occur in the space which essentially has vague boundary. In addition, this method can automatically decide the shape of the area in which such activities might occuer based on the concept of semi-supervised clustering. I applied this method to the artificial data and the actual data of pedestrians rendezvousing in Umeda Underground Mall and got good results.

Free Research Field

工学