朱容波,张静静,李媛丽,海梦婕,王德军.基于优化FOA-BPNN模型的脱贫时间预测[J].中南民族大学学报自然科学版,2018,(4):109-114
基于优化FOA-BPNN模型的脱贫时间预测
Time Prediction of Poverty Eradication with Optimized FOA-BPNN Model
  
DOI:10.12130/znmdzk.20180422
中文关键词: 精准扶贫  果蝇优化算法  脱贫时间预测  BP神经网络
英文关键词: targeted poverty alleviation  fluit fly optimization algorithm (FOA)  time prediction of povertyeradication  BPNN
基金项目:国家自然科学基金资助项目( 61772562) ; 湖北省技术创新专项软科学研究项目( 2018ADC150)
作者单位
朱容波,张静静,李媛丽,海梦婕,王德军 中南民族大学 计算机科学学院武汉 430075 
摘要点击次数: 127
全文下载次数: 108
中文摘要:
      针对精准扶贫缺乏有效的分析模型对扶贫的成效与脱贫时间进行准确刻画与定量分析问题,提出了基于 FOA-BPNN 贫困户脱贫时间预测模型.针对BP神经网络模型可能陷入局部最小的缺陷,引入果蝇优化算法,以BP 神经网络的预测误差作为适应度值,寻找最优的BP神经网络参数值,提高参数精度.由于标准果蝇优化算法的搜 索半径固定,可能导致后期局部寻优性能弱,提出了一种动态步长变更策略的DSFOA-BPNN模型,通过引入变速因 子与种群密度,将动态步长FOA算法与传统误差反向传播神经网络(BPNN) 结合,提高模型预测时间精度.在湖北 省某贫困地区 50000 条扶贫数据的基础上,通过实验表明:与BPNN和FOA-BPNN模型相比, 提出的 DSFOA-BPNN 模型预测精度分别提高了44%和11%.增量实验表明: 提出的DSFOA-BPNN模型更适用于精度预测.
英文摘要:
      For the lack of effective analytical models, a poverty eradication time prediction model is proposed based on FOA-BPNN, which aims to quantitatively analyse the effect of targeted poverty alleviation and precisely predict the time of getting rid of poverty. In order to overcome the shortage of BP neural network model which may fall into the local minimum, the fruit fly optimization algorithm ( FOA) is employed. FOA uses the error of training on BP neural network as fitness value to find the optimal BP neural network parameters and improve the accuracy of predicting. Due to the fixed search radius of the standard fruit fly optimization algorithm, it may lead to weaker performance in local optimization of FOA during the latter period of model's training. A DSFOA model using a dynamic step adjusting strategy is proposed, which introduces the shift factor and population density. And then BPNN is combined with DSFOA to improve the prediction accuracy of the model. Based on the 50, 000 samples of poverty alleviation in a poverty-stricken area in Hubei Province, experimental results show that the proposed DSFOA-BPNN model improves prediction accuracy by 44% and 11% respectively, compared with the BPNN and FOA-BPNN models. Meanwhile, the incremental experiment proves that the proposed DSFOA-BPNN model is more adaptive in prediction accuracy.
查看全文   查看/发表评论  下载PDF阅读器
关闭