摘要
为了解决网络舆情预警评价问题中样本数据的非线性和高维性等实际难题,提出以投影寻踪理论为基础,利用文化基因算法确定最佳投影方向,建立了基于文化基因算法的网络舆情评价的投影寻踪模型。将上述模型应用到2013年"新疆花帽"事件上,基于文化基因算法优化的投影寻踪模型可以很好的将各时间段的网络舆情进行划分等级,并能确定了各评价指标对网络舆情的影响。实例研究表明,基于文化基因算法的投影寻踪模型能有效地解决样本数据非线性、高维数等实际难点,有很好的实用性。
In order to solve the praetieal problem of nonlinear and high dimensions of sample data, which appeared on evaluation of online public opinion, the Memetie Algorithm and Projection Pursuit Theory were used to determine the optimal projection direction, and built a projection pursuit model of the evaluation of online public opin- ion under the Memetie Algorithm. This model was applied to ＂ Xinjiang Uighur cap ＂ incident in 2009, the model could be effective to grade network public opinion in the period of time, and could determine the influence of each e- valuation index of network public opinion. The calculation result shows that the Projection Pursuit model based on Memetic Algorithm can effectively overcome the practical difficulties of nonlinear, high dimension of the sample data, and has good practicability.
作者
李喆
LI Zhe(Center of Network and Information Technology, Xinjiang University, Urumqi Xinjiang 830046, Chin)
出处
《计算机仿真》
北大核心
2017年第4期391-395,共5页
Computer Simulation
关键词
状态转移算法
神经网络
滤波器
权值优化
State transition algorithm
Neural network
Filter
Weighted value optimization
