Modeling urban development using statistical preprocessing techniques and Artificial Neural Network: The case of Tehran metropolis

In late 2011, the world’s population surpassed the 7 billion and it is currently growing by an additional 78 million persons every year. Most of the future population growth will occur in developing countries, particularly in less developed countries (United Nations, 2011: 7). This increasing population pressure is leading to unregulated growth (Maithani, 2009: 364), Sprawl, horizontal spread of urban areas, rapid changes in land use and increasing trend of environmental degradation (Dewan and Yamaguchi, 2009: 390). In order to alleviate ,adverse effects of urbanization on the environment and maintain optimal ecosystem function (Fang et al., 2005: 295), spatial and temporal patterns of land use and land cover changes (LULC), and subsequently the factors affecting these changes(Serra et al., 2008: 190), are important in developing rational economic, social and environmental policies (Long et al., 2007: 351 Dewan and Yamaguchi, 2009: 390). Also, the advent of satellite images and geospatial technologies has paved new dimension for assessing and monitoring land use/ cover changes (Mussie et al., 2011: 2149). Encountered to these severe negative impacts, there is an urgent need for urban planners to develop predictive models of urbanization. These models not only provide an understanding of the urban growth process but also provide realizations of the numerous potential growth scenarios taken by urban area in future (Maithani, 2010: 36). Despite enormous researchers on urbanization, modeling using intelligent methods is still obscure issues such as the ordering importance and number of input data in this field is required. Thus, due to lack of a systematic method to find the best combination of input parameters, we have on the way we have introduced amethod to fill this gap. The present study was aimed to model development of metropolitan Tehran using Multi-layer Perceptron Neural Network model and Ordinary Least Squares regression (OLS) for preprocessing the input parameters to the model. 2- Methodology To satisfy end of the present study, at first, effective criteria in the urbanization process were collected from associated organizations, analyze, and prepared. Land use map for the study period (2006-1995) was extracted from the Landsat images and was improved using AutoCAD data and available map. Then, accuracy assessment of maps and change detection were performed and based on these changes, in order to avoid the trial and error method for choosing the best combination of input parameters to the model, using OLS pre-processing was performed according to criteria. Then, given to output method of OLS, the independent variables were chosen as input to the model. Finally, using Multi-layer Perceptron algorithm of Artificial Neural Network, transition potential modeling for each criterion undertaken, and the Markov chain method the land use map for the year 2017 was predicted. 3- Discussion In the present study, Ordinary Least Squares regression (OLS) and Multi-layer Perceptron algorithm of Neural Network were used to identify and improve our understanding of the social-economic, physical and land use forces that affecting urbanization, as well as finding the unequal impact of these factors and the most likely location for future urban development of metropolitan Tehran. Due to the large number and variety of factors affecting in the process and research results it can be notedthat RS and GIS technology allow us to generate and analyze large amounts of spatial and spatial data and output with a high degree of accuracy in the shortesttime. The advantage of this study compared to previous studies, is to use OLS method on input independent variables to the preprocessed model. Previous studies did not refer to preprocessing on input variables to the model, and if so, trial and error method is used. Among them, research conducted by Tayebi et al and Kamiab et al (1390 year) on predicted urbanization of Gorgan city and that Karam and Yaghob Nezhade Asl (1392 year) on thephysical development of Karaj city are substantial. Also, surveys field and remotely sensed results showed that the predictions compared to previous studies are closer to the ground realities and follows the available development trend. 4- Conclusion Results of the present study show that most urbanization trends of metropolitan Tehran for theprospect of the year 2017 will be concentrated in the eastern and western parts. These results indicate the validity of the model, fully consistent with the facts and can be implemented as a model in planning for the future prospects of metropolitan Tehran. However, to upgrade and increase the reliability of the model more efficiently, future studies can be used in a promising manner such as property, land prices and commercial centers, that due to restrictions available are not used here, and the urbanization process of metropolitan Tehran can be very influential. Keywords: modeling urban development, statistic preprocessing, least squares regression, neural network, RS, GIS. -United Nations, (2011), Department of Economic and Social Affairs - Population Division, World population prospects. -Maithani, S, (2009), A Neural Network based Urban Growth Model of an Indian City, J. Indian Soc. Remote Sensing, 37, 363–376. -Dewan, A.M, Yamaguchi, Y, (2009), Land use and land cover change in Greater Dhaka, Bangladesh: Using remote sensing to promote sustainable urbanization, Applied Geography, 29, 390–401. -Fang, Sh, Gertner, G.Z, Sun, Zh, and Anderson, A.A, (2005), The impact of interactions in thespatial simulation of the dynamics of urban sprawl, Landscape and Urban Planning, 73, 294–306. -Serra, P, Pons, X, Saur, D, (2008), Land-cover and land-use change in a Mediterranean landscape: A spatial analysis of driving forces integrating biophysical and human factors, Applied Geography, 28, 189–209. -Long, H, Tang, G, Li, X, and Heilig, G.K, (2007), Socio-economic driving forces of land-use change in Kunshan, the Yangtze River Delta economic area of China, Journal of Environmental Management, 83, 351–364. -Maithani, S, (2010), Application of Cellular Automata and GIS Techniques in Urban Growth Modelling: A New Perspective, Institute of Town Planners, India Journal, 7, 36 – 49. -Mussie, G, Tewolde, Cabral, P, (2011),Urban Sprawl Analysis and Modeling in Asmara, Eritrea, Remote Sensing, 3, 2148-2165.