Evaluation of the urban development patterns of Beijing from 1988 to 2007 by time series classification approach

Urban development produces significant impacts in society and its environment. After implementing the reform and opening policies, Beijing has experienced rapid urban development during recent decades. This paper evaluates the urban development patterns of Beijing from 1988 to 2007 by using time series classification approach. The study area is divided into cells with 30m spatial resolution. Datasets on urban development of Beijing in different time points, including 1988, 1995, 1999, 2002 and 2007 is collected. For each time point, six driving factors of urbanization are chosen, including land use type, distance to traffic network, distance to school, distance to market, distance to city center, and slope. Based on Analytic Hierarchy Process (AHP) approach, relative weights of these factors contributing to the urbanization level are calculated. An index describing the urbanization level of each cell at each time point in Beijing is generated by synthesizing the weights of the aforementioned factors. A curve which shows the development pattern over time is formed for each cell. Self-Organizing Map (SOM) and K-means clustering algorithm are introduced to extract prototype curves from the time series dataset. Each cell of Beijing is recognized and classified to certain development pattern. Therefore, a classification map contain seven development patterns is formed. By analyzing this map, the spatial-temporal patterns of urban development of Beijing are discussed.