基于多维关联规则兴趣度的问卷调查规则提取

Extraction of Multidimensional Association Rules

基于能同时处理多个属性间关联关系的多维关联规则算法,对大学生社交网络行为习惯的调查问卷进行研究分析,发现依靠支持度和置信度的关联规则算法有时会产生误导性的结果。针对关联规则存在的这一问题,给出了带有负向的关联规则兴趣度的解决办法,并发现兴趣度规则中减少关联规则计算量的性质,可极大提高了多维关联规则兴趣度算法在规则提取中的效率。实验结果表明,负向的关联规则置信度强于正向的关联规则置信度,引入兴趣度的多维关联规则算法的准确度更高。

Applying the multidimensional association rule algorithm which can deal with the incidence relation between many attributes simultaneously, in this paper, it is found that the association rule algorithm depending on support degree and confidence coefficient may cause some misleading results according to the questionnaires on the college students' safety awareness of suffering Internet. To solve the problem of association rules, a method with a negative association rules with interest degree is proposed. The amount of calculation and found that properties of association rules to reduce interest degrees of rules, which can improve the efficiency of multidimensional association rule interestingness in rules extraction algorithm. The experiment result shows that the negative association rules are stronger than positive association rules in confidence, and the multidimensional association rules algorithm with interestingness are more accurate.