【摘要】 针对农村经济预测能力不足、预测方法不适用、预测准确度差等问题,提出使用神经网络中的GRNN方法,通过农业总产值及粮食等农作物产量,对农业机械动力及农村用电量需求进行预测。GRNN模型具有样本数据少、非线性映射能力强、易于进行多元、多目标预测的特点。同时,该模型人为调节的参数少,学习、训练时间短,使用方便,很适合农村管理人员的经济预测需求。论文的实证案例经建模、训练、仿真,选取了误差小、拟合度高的模型并应用于农村经济预测。研究结果表明,前向神经网络(包括GRNN)方法适合广大农村经济管理应用,较具推广价值。类似方法的推广应用必将使广大农村"只有统计,没有预测"的现实得以改观。更多还原

【Abstract】 In view of the rural economy predictive ability insufficiency, the forecast technique are not suitable, forecast accuracy difference, the neural netuork’s GRNN method be proposed, through total agricultural output value and grain crop yield and so on, carries on the fore- cast to the farm machinery power and the countryside electricity consumption demand. The GRNN model has the sampled data to be few, non-linear mapping ability strong, easy to carry on multi-dimensional, the multi-objective forecast characteristic. At the same time, this model artificial adjustment’s parameter are few, the study, the training time are short, easy to operate, very suitable countryside administrative personnels’ economic projection demand. The paper real diagnosis case after the modelling, the training, the simulation, selected the error to be small, the fitting high model and applied in the rural economy forecast. The findings indicated that the forward neural network (including GRNN) method suits the general rural economy management application, has the value to promote. The similar method’s promoted application will certainly to cause the general countrysides "only then to count, had not forecast" the reality can have a new look.更多还原