| Cluster analysis is a useful mathematical method that has been used to examine the differences or similarities between gardens (Xu Y. , 2015). This investigation aims to test the differences and similarities between traditional Japanese gardens in Kyoto, the classical Chinese gardens in Suzhou, and the modern Chinese gardens in Xiamen, by comparing the design elements and design principles of theose gardens. A hundred and thirty-four variables are selected based upon a literature review and the author's personal experience in Kyoto, where the first seventy-five variables are adopted from earlier garden research by Yiwen Xu (Xu Y. , 2015). After collecting the variables and applying the Principal Component Analysis by the software SAS, a group of corresponding eigenvalues are generated. According to the Principal Component Analysis, the first two principal components covered 63.81 percent of the sample variance. The first and second principal component together divided the gardens into three groups:; the first principal component indicated the similarities between theof traditional Japanese gardens and classical Chinese gardens; the second principal component indicatedmplied the similarities betweenof traditional Japanese gardens and modern Chinese gardens. Therefore, the final result is a group of three two-dimensional scatter graphs, where each point represents a garden's character on two coordinate axes, and the dimensions are generated from the meaningful eigenvalues. |
