林荫, 杨长春. 基于改进极限学习机的微信热点预测[J]. 微电子学与计算机, 2017, 34(5): 123-127.
引用本文: 林荫, 杨长春. 基于改进极限学习机的微信热点预测[J]. 微电子学与计算机, 2017, 34(5): 123-127.
LIN Yin, YANG Chang-chun. Weixin Hot Spot Prediction by Using Modified Extreme Learning Machine[J]. Microelectronics & Computer, 2017, 34(5): 123-127.
Citation: LIN Yin, YANG Chang-chun. Weixin Hot Spot Prediction by Using Modified Extreme Learning Machine[J]. Microelectronics & Computer, 2017, 34(5): 123-127.

基于改进极限学习机的微信热点预测

Weixin Hot Spot Prediction by Using Modified Extreme Learning Machine

  • 摘要: 为了提高微信热点预测的准确性, 更好描述微信热点的变化趋势, 针对当前微信热点预测模型存在的一些局限性, 设计了改进极限学习机的微信热点预测模型.首先采集微信热点的历史数据, 构建微信热点的时间序列学习样本, 然后采用极限学习机对微信热点样本进行训练, 并对标准极限学习机进行改进, 提升其学习能力, 建立了微信热点的预测模型, 最后采用具体微信热点数据进行验证性测试.测试结果表明, 改进极限学习机可以获得高精度的微信热点预测结果, 而且微信热点的建模效率要高于其他微信热点预测模型, 具有良好的实际应用价值.

     

    Abstract: In order to improve the accuracy of the weixin hot spot prediction, describe weixin hot trends, the current weixin hot forecast model has some limitations, a weixin hot prediction model based on improved extreme learning machine is proposed. Firstly, historical data are gathered for weixin hot, and time series hot study sample of weixin establisheds, and secondly, extreme learning machine was trained on weixin hot samples while standard extreme learning machine learning is modified to improve learning ability and prediction models of weixin hot is test by weixin hot data test. The test results show that modified extreme learning machine can get high accurate prediction results of weixin hot, and the modeling efficiency is higher than other models, it has good practical application value.

     

/

返回文章
返回