基于遗传算法优化BP神经网络的中国企业OFDI投资额预测模型  

Investment Prediction Model of Chinese Firms'OFDI Based on BP NeuralNetwork Optimized by Genetic Algorithm

作　　者：刘峻杉 张磊[1,2] 尹寓 LIU Jun-shan;ZHANG Lei;YIN Yu(College of Computer Science,Sichuan University,Chengdu 610065;Big Data Analysis and Fusion Application Technology Engineering Laboratory of Sichuan Province,Chengdu 610065)

机构地区：[1]四川大学计算机学院,成都610065 [2]四川省大数据分析与融合应用技术工程实验室,成都610065

出　　处：《现代计算机》2021年第7期29-36,共8页Modern Computer

基　　金：四川省科技厅应用基础研究项目(No.2019YJ0110);四川省科技服务业示范项目(No.2019GFW167)。

摘　　要：中国企业对外直接投资是一个复杂且具有挑战性的问题,往往受到多个层面的影响,是一个多因素决策的过程。传统的金融统计方法难以解释多变量之间的关系。为了有效解决这个问题,将金融背景与计算机方法有机结合,采用自组织映射与模糊C均值对2004-2015年参与对外直接投资的1000多家中国企业的投资区位进行了划分,分别建立基于遗传算法优化BP神经网络的投资额预测模型。并与传统的BP神经网络进行比较,验证该模型的有效性,为中国企业对外直接投资提供新思路。Chinese firms’outward foreign direct investment is a complex and challenging problem,which is often affected by multiple levels and is a multi-factor decision-making process.Traditional financial statistical methods are difficult to explain the relationship between variables.In order to effectively solve this problem,the financial background and computer methods are combined organically,and the investment lo⁃cation of more than 1000 Chinese firms participating in foreign direct investment from 2004 to 2015 is divided by using self-organizing maps and fuzzy C-means,and the investment prediction model based on genetic algorithm optimized BP neural network are established re⁃spectively.Compared with the traditional BP neural network,the validity of the model is verified,which provides a new idea for Chinese firms to invest abroad.

关 键 词：企业对外投资 自组织映射 模糊C均值 遗传算法 BP神经网络 

分 类 号：F125[经济管理—世界经济] F279.2[自动化与计算机技术—控制理论与控制工程] TP18[自动化与计算机技术—控制科学与工程]