The false positive and false negative rates are improved using user feedback on the misclassified information with incremental learning to dynamically adjust the rule weights.

• 增加用户与服务器间的交互,收集用户反馈的错判信息,根据反馈信息用增量学习动态调整规则权重。