DETERMINATION OF URBAN CLASSIFICATION USING CART
GURAGAI, MANOJ KUMAR
MetadataShow full item record
Urbanization level is a valuable indicator for projections of some global trends. Now, urbanization levels may be based on unreliable data obtained from the sources. This thesis will study a newer approach of classification of urban areas through the machine learning technique. Generally, the time series analysis of urban and rural populations of a nation is done using manual calculation or using statistical tools. The time consuming, tedious and cumbersome approach may slightly solve using machine learning. The classification and regression tree (CART) algorithm is used to classify the data consisting of both quantitative and categorical variables. Cart is a special type of data mining algorithm and is to be used for inferring knowledge residing in the given data. The CART algorithm is used here as the machine learning algorithm to determine the classification of urban areas as defined for our nation. The classified urban areas are forming the decision tree according to the data provided by Central Bureau of Statistics (CBS), Ministry of federal affairs and local development (MOFALD), and Municipality Association of Nepal. The classification shows the tree on the basis of governmental classification as Metropolitan city, Sub-Metropolitan city and Municipality. This approach is also helping to classify other VDC’s developing to meet the criterion needing to become the urban areas. So, it is extendible to train and find the new urban areas.