计算机技术与发展
計算機技術與髮展
계산궤기술여발전
COMPUTER TECHNOLOGY AND DEVELOPMENT
2014年
7期
192-195
,共4页
遗传算法%自适应%神经网络%客户分类
遺傳算法%自適應%神經網絡%客戶分類
유전산법%자괄응%신경망락%객호분류
genetic algorithm%adaptive%neural network%customer classification
银行产品的营销行为都是针对广大客户的。若能提前分辨出哪些是优质客户,再为其定制合理的营销策略,那银行就能获得更大的竞争力。文中将遗传算法与BP神经网络结合用于对银行客户分类进而预测客户是否会购买银行产品。该方法有效地克服了BP神经网络容易陷入局部极小值和收敛速度慢的问题,并且针对其中遗传算法的计算时间和精度问题提出了一种新的自适应遗传算法。实验结果表明,基于这种自适应的遗传神经网络的方法用更短的计算时间达到了更高的预测精度,可以准确地为银行客户分类。
銀行產品的營銷行為都是針對廣大客戶的。若能提前分辨齣哪些是優質客戶,再為其定製閤理的營銷策略,那銀行就能穫得更大的競爭力。文中將遺傳算法與BP神經網絡結閤用于對銀行客戶分類進而預測客戶是否會購買銀行產品。該方法有效地剋服瞭BP神經網絡容易陷入跼部極小值和收斂速度慢的問題,併且針對其中遺傳算法的計算時間和精度問題提齣瞭一種新的自適應遺傳算法。實驗結果錶明,基于這種自適應的遺傳神經網絡的方法用更短的計算時間達到瞭更高的預測精度,可以準確地為銀行客戶分類。
은행산품적영소행위도시침대엄대객호적。약능제전분변출나사시우질객호,재위기정제합리적영소책략,나은행취능획득경대적경쟁력。문중장유전산법여BP신경망락결합용우대은행객호분류진이예측객호시부회구매은행산품。해방법유효지극복료BP신경망락용역함입국부겁소치화수렴속도만적문제,병차침대기중유전산법적계산시간화정도문제제출료일충신적자괄응유전산법。실험결과표명,기우저충자괄응적유전신경망락적방법용경단적계산시간체도료경고적예측정도,가이준학지위은행객호분류。
The products in bank marketing are faced to the majority of customers. If tell in which are high-quality customers in advance and then develop reasonable marketing strategy for them,bank will be able to achieve greater competitiveness. It combines genetic algo-rithm with BP network for bank customers classification to predict whether the customers will buy the bank marketing products. It can ef-fectively overcome the shortcomings of BP network,such as trapping to the local minimum and slowness in training speed. Aiming at the computation time and accuracy of genetic algorithm,a new adaptive GA-BP algorithm is proposed. Experimental results show that the a-daptive GA-BP algorithm can reach a higher prediction accuracy with a shorter calculation time and it can classify bank customers accu-rately.