Classification of Economic Characteristics based on Klassen Typology using Multilayer Perceptron Neural Network

摘要

Urban development in Java is undergone rapidly during the last two decades. One factor driving urban development is economic characteristics. The autonomy policy in Indonesia might bring economic disparity among districts in Java. This study aims to develop the classification of economic characteristics based on Klassen typology in Java using multilayer perceptron neural networks. Two aspects will be considered in building model i.e. demography and physical. Economics data, demographics data, and physical data was acquired. The economy characteristics of districts in Java was classified based on Klassen typology using economic data. We designed the neural networks based on the multilayer perceptron (MLP) to classify the economic characteristics. In the framework of the neural networks, the physical data and demographic data were used as the input of the networks. The economy characteristics based on Klassen typology then used as the output of the networks. The parameters were varied to assess the performance. In the experiments, the results indicated that our proposed method works efficiently to classify economic characteristics based on Klassen typology in almost regions.