基于灰色-马尔可夫藕合模型的区域建设用地需求预测

Prediction of construction lands by Gray-Markov models

建设用地需求的中长期准确预测是当前土地资源规划和利用的重点和难点。本文将灰色和马尔可夫两种预测模型相结合,建立灰色-马尔可夫藕合模型,以广东省揭阳市为例,对1997到2010年间主要建设用地规模进行了需求预测,同时与其它几种常规的预测方法进行了比较。结果表明,人文驱动力相关预测法误差相对较大,其中国内生产总值(GDP)相关法准确度要优于人口相关法。趋势线法计算简单,预测准确度也相对较高,但只适于短期预测。灰色预测法精确度较高,计算需要模型的支持、专业性强。灰色-马尔可夫预测法由于是对灰色预测的偏离值进行了纠正,因而准确度最高。尽管该方法需要大量数据的支持,计算量大,但仍是值得推荐的高准确度的区域建设用地的专业预测方法。

Medium and long-term prediction of construction land is the key of planning and uses of land resources. We combined both Gray and Markov model to Gray-Markov model. Taking Jieyang city as example, we predicted the main construction land during 1997-2010 with Gray-Markov model. Meanwhile, we compared our prediction results with other conventional forecasting methods. Out results indicate that the method of human driving force has obvious prediction errors. Correlation analysis method of gross domestic product （GDP） has higher accuracy than method of population correlation analysis. Although the trend method is simple and relatively accurate, it is only suitable for short-tenn forecasts. Furthermore, Gray prediction method is accurate, but it need support of calculation model or requires professional knowledge. In comparison, Gray-Markov prediction, as a corrected gray method, provides the highest accuracy among all method above. Despite a large amount of data and calculation required, this method is demonstrated as an accurate prediction method of the regional construction lands.