基于改进粒子群优化神经网络的房地产市场预测
发布时间：2019-05-13 19:38
【摘要】：针对粒子群优化算法精度不高、容易陷入局部最优、难以满足房地产市场形势需求的问题,提出一种改进粒子群优化神经网络,并应用于房地产市场预测中,该算法将混沌引入粒子群优化神经网络算法权重和阈值的初始化与更新的过程,提高了初始样本的质量,减轻了局部极值现象,提高了算法的全局搜索能力,同时设置了躲避因子,使粒子一定程度上离开偏离真实值的区域。研究结果表明,提出的改进算法可以提高粒子群优化神经网络权重和阈值的准确性。
[Abstract]:In order to solve the problem that the accuracy of particle swarm optimization algorithm is not high, it is easy to fall into local optimization, and it is difficult to meet the needs of real estate market situation, an improved particle swarm optimization neural network is proposed and applied to the prediction of real estate market. Chaos is introduced into the initialization and updating process of weight and threshold of particle swarm optimization neural network algorithm, which improves the quality of initial samples, reduces the local extremum phenomenon and improves the global search ability of the algorithm. At the same time, the avoidance factor is set to make the particles leave the region that deviates from the real value to a certain extent. The results show that the improved algorithm can improve the accuracy of particle swarm optimization neural network weight and threshold.
【作者单位】： 山东城市建设职业学院;南京理工大学自动化学院;
【基金】：国家自然科学基金资助项目(61104109) 教育部博士点基金资助项目(20113219110027) 江苏省自然科学基金资助项目(BK2011703) 江苏省科技支撑与自主创新基金资助项目(BE2012178)
【分类号】：F299.23;TP18
[Abstract]:In order to solve the problem that the accuracy of particle swarm optimization algorithm is not high, it is easy to fall into local optimization, and it is difficult to meet the needs of real estate market situation, an improved particle swarm optimization neural network is proposed and applied to the prediction of real estate market. Chaos is introduced into the initialization and updating process of weight and threshold of particle swarm optimization neural network algorithm, which improves the quality of initial samples, reduces the local extremum phenomenon and improves the global search ability of the algorithm. At the same time, the avoidance factor is set to make the particles leave the region that deviates from the real value to a certain extent. The results show that the improved algorithm can improve the accuracy of particle swarm optimization neural network weight and threshold.
【作者单位】： 山东城市建设职业学院;南京理工大学自动化学院;
【基金】：国家自然科学基金资助项目(61104109) 教育部博士点基金资助项目(20113219110027) 江苏省自然科学基金资助项目(BK2011703) 江苏省科技支撑与自主创新基金资助项目(BE2012178)
【分类号】：F299.23;TP18
【相似文献】
相关期刊论文 前10条
1 李天牧,李学群;基于神经网络的自动特征抽取[J];云南大学学报(自然科学版);1991年03期
2 kelly MF;邰常峰;;神经网络在肌电信号分析中的应用[J];国际生物医学工程杂志;1991年06期
3 汤新梁;;自由手写体数字识别的神经网方法[J];东南大学学报(自然科学版);1991年03期
4 卢科学;郭美义;;平移不变模式变换与神经网络[J];西安电子科技大学学报;1992年03期
5 王继成,吕维雪;基于神经网络的逻辑计算[J];计算机工程与设计;1993年02期
6 丁德恒;吕维雪;;基于神经网络的关系模式匹配方法[J];计算机科学;1993年01期
7 夏又生，，吴新余;解框形约束最小二乘问题的神经网络[J];南京邮电学院学报;1994年03期
8 沈琴婉;黄五群;陈天];张延p
本文编号：2476136
本文链接：https://www.wllwen.com/jingjilunwen/fangdichanjingjilunwen/2476136.html
