论文部分内容阅读
[目的/意义]微博在当前信息传播中起着重要作用,为有效预测微博热点及舆情导控,建立实时线性预警模型。[方法/过程]将采集的指标进行缺失值和异常值的处理后,对微博话题热度与大V影响力因子进行因子分析与逐步回归的比较,筛选出公共影响因子;再对其加权,探索不同权重调节因子下的最佳定量公式;用此公式每次输入当前时刻起前3小时的数据,预测当前时刻起后30分钟的加权值对应的话题词,每隔10分钟重新更新一遍参数。[结果/结论]实验证明该预测模型能大大降低数据采集解析和预测时间,保持较好的准确率,并可通过选择合适的阈值,进一步提升精确度。
[Purpose / Significance] Weibo plays an important role in the current information dissemination. In order to effectively forecast the hotspot and public opinion in Weibo, we established a real-time linear early warning model. [Method / Process] After the missing index and abnormal value are processed, the factor of the microblog topic heat and the influence factor of large V are compared by factor analysis and stepwise regression, and the public influencing factors are screened out. Secondly, Explore the best quantitative formula under different weight adjustment factors; input the words of the first 3 hours from the current moment by using this formula, predict the topic words corresponding to the weighted values of 30 minutes after the current moment, and update the parameters again every 10 minutes. [Result / Conclusion] The experiment proves that this prediction model can greatly reduce the data collection resolution and prediction time, maintain a good accuracy, and can further improve the accuracy by selecting a suitable threshold.