CAJ | 학술논문

提出了利用径向基神经网络对农机数量预测的方法,通过选取合适的训练参数,径向基网络能够得到满足要求的预测结果,农机总动力、拖拉机数量、农具数量的误差平方和分别为0.0056、0.0470、0.2713.利用测试集对网络进行测试,农机总动力预测值与真实值的误差最大为-7.17%,最小为0.22209%.研究结果表明,径向基神经网络能有效提高预测精确度,较好地预测非线性条件下的农机数量,证明了实验方法的有效性和可行性,为人工神经网络在农业机械化的应用提供了一个新的途径.
제출료이용경향기신경망락대농궤수량예측적방법,통과선취합괄적훈련삼수,경향기망락능구득도만족요구적예측결과,농궤총동력、타랍궤수량、농구수량적오차평방화분별위0.0056、0.0470、0.2713.이용측시집대망락진행측시,농궤총동력예측치여진실치적오차최대위-7.17%,최소위0.22209%.연구결과표명,경향기신경망락능유효제고예측정학도,교호지예측비선성조건하적농궤수량,증명료실험방법적유효성화가행성,위인공신경망락재농업궤계화적응용제공료일개신적도경.
@@@@Put forward a method to predict the number of agricultural machinery by using RBF NN. By choosing proper training parameters, RBF network can get the predicted results of meeting the requirements. The error sum of squares of agricultural machinery total power, trac-tor number and farm implements number is 0.0056, 0.0470, 0.271, to test the network Using test sets. The maximum error of agricultural machinery total power between predicted value and real value is-7.17%. The minimum is 0.22209%.The research results indicate that RBF NN can effectively increase the prediction accuracy. The number of agricultural machinery under the condition of nonlinear can be well pre-dicting. Experimental method is proved to be feasible and effective. It provides a new way for the application of artificial neural network in agricultural mechanization.