An Application of Spectral Clustering on Vector Clustering

• Main Authors: CHAN, BO NING ANNABEL, 陳寶寧
• Other Authors: LEE, MONG-HONG
• Format: Others
• Language: en_US
• Published: 2019
• Online Access: http://ndltd.ncl.edu.tw/handle/tt9h6w

碩士 === 國立臺北大學 === 統計學系 === 107 === In a decade, the demand for services about home delivery and online cars calling rapidly increase due to the development and improvement of technology and e-commerce. The growth of these industries has made route planning more important. Clustering locations efficiently could reduce the operating cost and improve environmental problems caused by carbon dioxide emissions from vehicles. In the example of a company that provides pick-up service to disable people by rehabilitation bus, they usually cluster the locations and orders by self experiences. However, it would be time-consuming and inefficient if the amount of data is big, and it's a waste of manpower to the company. Therefore, spectral clustering would be used in this paper to deal with datasets with vector structure, and implement clustering by seeking latent features. Finally, it will compare with the results using k-means.