灰色理论和人工神经网络森林资源预测方法的对比研究

以森林资源一类清查数据为基础,采用灰色系统理论、人工神经网络两种方法对山西省森林资源的发展趋势进行了预测,并对两种方法的优缺点进行了对比研究。结果表明：人工神经网络的相对误差低于灰色理论,拟合性高。采用人工神经网络预测森林资源变化是一种较好的方法,且对数据有较好的适应能力;人工神经网络更适合于短期预测,而灰色理论适合于中长期的预测。在进行森林资源预测时,应根据不同目的把两种方法结合使用。

Comparative Study of Predictive Method of Two Forest Resources Based on Gray Theory and the Artificial Neural Network

Based on the forest resources continuous check data, the development trend of forest resources of Shanxi Province is predicted by using Gray Model （GM） and Artificial Neural Network （ANN）. The results of comparative study about the two prediction methods show that, the relative error of ANN is lower than GM, so its regression property is comparatively high. The model has relative good prediction accuracy and feasibility for data than the gray theory GM. But the Artificial Neural Network is more adapted to short-term prediction, Gray Model is suited to middle-long period prediction of forest resources. So we should integrate each other in the light of particular purpose.