Resources Science ›› 2015, Vol. 37 ›› Issue (9): 1797-1806.

• Orginal Article • Previous Articles     Next Articles

A case-based cellular automation model for simulating changes in rural residential areas

GONG Jian1,2(), YANG Jianxin1(), LI Yafang1   

  1. 1. School of public management,China university of Geosciences (Wuhan),Wuhan 430074,China
    2. Key Laboratory of the Ministry of Land and Resources Law Evaluation,Wuhan 430074,China
  • Received:2014-09-13 Revised:2015-01-26 Online:2015-09-25 Published:2015-09-25

Abstract:

As at 2011 China had a rural population of 657 million people,accounting for 48.7% of the total population. The rural residential area,as a gathering place for rural people,still accounts for the vast proportion of urban and rural residential land in China. The study of rural residential area has mostly focused on rural residential land consolidation,spatial distribution optimization and the driving forces of evolution;however,simulating the spatial distribution changes of rural residential area remains rare. Here,we applied cellular automata modeling to simulate rural residential area changes,gained cellular transformation rules using case based reasoning methods based on analogical reasoning principle of k-Nearest Neighbor similarity,determined the new case state at reasoning only by accounting similarities of cases with states already changed,drive variable standard deviations to decide the variable weight,applied the ROC method to determine the optimum valve of the only parameter k of the model,applied the Markov method to take total number control over the model,and used the improved Kappa coefficient to verify the accuracy of simulation results. Rural residential area from 1991 to 2004 was the model training data;calibration of model parameters was done with data for changes from 2004 to 2009; rural residential area distribution situation of 2009 was simulated by contrasting the actual situation for the study area in 2009. The total percentage accuracy of simulation results was 96.68% and the improved Kappa coefficient value was 0.93. The results show that the cellular automata model based on case based reasoning has a higher simulation accuracy,which can be used in simulation and analysis of changes in rural residential areas.

Key words: case-based reasoning, cellular automation, Kappa coefficient, rural residential