| 基于MQPSO-LSSVM的微博热点话题预测 |
| |
| 引用本文： | 符保龙.基于MQPSO-LSSVM的微博热点话题预测[J].柳州师专学报,2013(6):117-120,113. |
| |
| 作者姓名： | 符保龙 |
| |
| 作者单位： | 柳州职业技术学院,广西柳州545006 |
| |
| 基金项目： | 广西教育厅科研项目基金资助（201106LX745,201204LX593）. |
| |
| 摘 要： | 微博热点话题预测是一类小样本、不确定性的复杂预测问题,传统线性方法不能刻画微博热点话题的变化规律,神经网络存在过拟合、泛化能力不强等缺陷.为了提高微博热点话题的预测精度,提出了一种改进量子粒子群(QPSO)算法优化LSSVM的微博热点话题预测模型(MQPSO-LSSVM).首先采用MQPSO算法优化LSSVM的参数,然后将优化后的LSSVM对微博热点话题变化趋势进行建模,最后选取具体微博热点话题数据进行仿真实验.实验结果表明,MQPSO-LSSVM提高了微博热点话题的预测精度,预测结果具有一定实用价值.
|
| 关 键 词： | 微博热点话题 量子粒子群算法 参数优化 最小二乘支持向量机 |
Micro Blogging Hot Topic Prediction Based on LSSVM Optimized by Improved MQPSO Algorithm |
| |
| FU Baolong.Micro Blogging Hot Topic Prediction Based on LSSVM Optimized by Improved MQPSO Algorithm[J].Journal of Liuzhou Teachers College,2013(6):117-120,113. |
| |
| Authors: | FU Baolong |
| |
| Institution: | FU Baolong (Liuzhou Vocational and Technological College, Guangxi, Liuzhou, 545006 China) |
| |
| Abstract: | Miero-blog hot topic prediction is small sample, complex prediction problem, traditional linear methods can not describe the characterize of the micro-blog hot, and neural network has the defects of over-fitting, the generalization ability is not strong. In order to improve the prediction accuracy of micro-blog hot topic, this paper proposes a novel prediction model （MQPSO-LSSVM） based on improved quantum behaved particle swarm optimization （QPSO） and least squares support vector machine （LSSVM）. Firstly, parameters of LSSVM are optimized by the MQPSO algorithm, and then the optimized LSSVM is used to prediction miero-blog hot topic change, finally the simulation experiment is carried out on micro-blog hot topic data.The experimental results show that MQPSO-LSSVM improves the prediction accuracy of micro-blog hot topic, and the prediction results have certain practical value. |
| |
| Keywords: | micro-blog hot topic quantum behaved particle swarm optimization parameters optimization least squares support vector machine |
| 本文献已被 维普 等数据库收录！ |
|
