为了预测论坛舆情及其动态演变趋势,基于多时间序列的关联分析,集中分析了论坛中3个量的时间序列之间的关联规则:活跃者之间的关系强度的时间序列、坚定支持者人数的时间序列以及坚定支持者成员的变化频度的时间序列。然后给出了一种新...为了预测论坛舆情及其动态演变趋势,基于多时间序列的关联分析,集中分析了论坛中3个量的时间序列之间的关联规则:活跃者之间的关系强度的时间序列、坚定支持者人数的时间序列以及坚定支持者成员的变化频度的时间序列。然后给出了一种新的基于多时间序列关联分析的论坛舆情预测算法(Forum sentiment trend prediction based on multi time series association rule analysis,TPMTSA),并在真实数据集和拟合数据集上进行了大量的实验。结果表明:TPMTSA算法具有有效性和较高的运行效率。研究结果可用于论坛舆情预警监控。展开更多
Incompleteness of information about objects may be the greatest obstruct to performing induction learning from examples. In this paper, the concept of limited-non-symmetric similarity relation is used to formulate a n...Incompleteness of information about objects may be the greatest obstruct to performing induction learning from examples. In this paper, the concept of limited-non-symmetric similarity relation is used to formulate a new definition of approximation to an incomplete information system. With the new definition of approximation to an object set and the concept of attribute value pair, rough-setsbased methodology for certain rule acquisition in an incomplete information system is developed. The algorithm can deal with incomplete data directly and does not require changing the size of the original incomplete system. Experiments show that the algorithm provides precise and simple certain decision rules and is not affected by the missing values.展开更多
文摘为了预测论坛舆情及其动态演变趋势,基于多时间序列的关联分析,集中分析了论坛中3个量的时间序列之间的关联规则:活跃者之间的关系强度的时间序列、坚定支持者人数的时间序列以及坚定支持者成员的变化频度的时间序列。然后给出了一种新的基于多时间序列关联分析的论坛舆情预测算法(Forum sentiment trend prediction based on multi time series association rule analysis,TPMTSA),并在真实数据集和拟合数据集上进行了大量的实验。结果表明:TPMTSA算法具有有效性和较高的运行效率。研究结果可用于论坛舆情预警监控。
文摘Incompleteness of information about objects may be the greatest obstruct to performing induction learning from examples. In this paper, the concept of limited-non-symmetric similarity relation is used to formulate a new definition of approximation to an incomplete information system. With the new definition of approximation to an object set and the concept of attribute value pair, rough-setsbased methodology for certain rule acquisition in an incomplete information system is developed. The algorithm can deal with incomplete data directly and does not require changing the size of the original incomplete system. Experiments show that the algorithm provides precise and simple certain decision rules and is not affected by the missing values.