CAJ | 학술논문

提出了一种融合遗传算法与BP神经网络的城镇基准地价评估模型。该模型以正方形网格作为基本评价单元,以遗传算法优化BP神经网络连接权值,以BP神经网络训练样本数据,实现其他网格地价的模拟与预测。结果表明,遗传算法可以有效增强BP神经网络快速学习网络权值能力和克服收敛速度慢、易陷入局部极小值等问题,BP神经网络的网络训练学习与自适应性符合处于动态变化的土地价格发展要求,呈现出广阔的应用前景。
제출료일충융합유전산법여BP신경망락적성진기준지개평고모형。해모형이정방형망격작위기본평개단원,이유전산법우화BP신경망락련접권치,이BP신경망락훈련양본수거,실현기타망격지개적모의여예측。결과표명,유전산법가이유효증강BP신경망락쾌속학습망락권치능력화극복수렴속도만、역함입국부겁소치등문제,BP신경망락적망락훈련학습여자괄응성부합처우동태변화적토지개격발전요구,정현출엄활적응용전경。
A combination of genetic algorithm and BP neural network evaluation model of urban standard land price is presented.Based on a square grid as the basic evaluation unit,the genetic algorithm to optimize BP neural network connection weights to BP neural network training sample data,the model realizes other land grid simulation and prediction.The results show that genetic algorithm can effectively enhance the BP neural network ability to quickly learn network weights and to overcome slow convergence,and is easy to fall into local minimum problems,BP neural network learning and adaptive network training in keeping with the dynamic changes in land prices development requirements,which has a broad applied prospects.