摘要
提出了一种新的基于概率模型的排序优化方法 .利用贝叶斯规则 ,结合各组成系统平均执行性能的信息 ,推导出一种新的相关度计算公式 ,较好地解决了结果融合中相关度规范化和均衡化的问题 .经实验验证 ,该方法对结果进行了最优化排序 ,其实际执行性能超出了现有的任何一个组成系统的性能 .
The paper put forward a new optimal scheme based on a probability model. By using the Bayes rule and combining the information of the average performance of the component systems, a new relation degree formula was deduced and the problems on relational degree standardization and equilibrium were solved. The experiment results showed the new system provided the optimal ranking.
出处
《华中科技大学学报（自然科学版）》
EI
CAS
CSCD
北大核心
2003年第3期49-51,共3页
Journal of Huazhong University of Science and Technology(Natural Science Edition)
基金
国家高性能计算基金资助项目 ( 993 1 9)
关键词
元搜索引擎
概率模型
结果优化排序
排序融合
metasearch engine
probability model
results optimal ranking
ranking and fusing
