一种面向动态偏好的交互式多目标查询

发布时间：2018-06-26 04:22

本文选题：多目标查询 + 动态偏好　；参考：《小型微型计算机系统》2017年12期

【摘要】：同一问题的迭代查询之间存在关联,且不同查询间用户偏好动态变化.但现有查询算法只是将不同查询独立处理或默认查询间偏好固定不变,本文提出一种可有效利用查询间关联关系且支持动态偏好的查询处理算法IMQD;定义一种结果集质量度量方法,取值在[0,1]区间内且取值越大越优.IMQD算法旨在使结果集质量更高,将查询处理过程分为三个步骤:首先确定初始候选集;然后通过用户交互获取调整后的偏好阈值,对当前候选集进行调整,使结果集质量取值更大;用户根据当前结果集质量判定是否终止查询.在模拟和真实数据集上验证IMQD算法可提升结果集质量且拥有良好交互性能.一旦用户偏好发生变化,算法可在毫秒级内完成对结果集的自适应调整.
[Abstract]:There is a correlation between iterative queries of the same problem and dynamic changes of user preferences among different queries. However, the existing query algorithms only fix the preferences between different query independent processing or default queries. In this paper, a query processing algorithm IMQD, which can effectively utilize the correlation between queries and support dynamic preference, is proposed, and a method of measuring the quality of result set is defined, in which the value is in the range of [0 ~ 1] and the larger the value is, the better. IMQD algorithm aims at making the result set better. The process of query processing is divided into three steps: first, the initial candidate set is determined; secondly, the adjusted preference threshold is obtained through user interaction, and the current candidate set is adjusted to make the quality of the result set greater. The user determines whether to terminate the query according to the quality of the current result set. The IMQD algorithm can improve the quality of the result set and has good interaction performance. Once the user preference changes, the algorithm can adjust the result set in milliseconds.
【作者单位】：南京航空航天大学计算机技术与科学学院;
【基金】：国家自然科学基金项目(61373015)资助
【分类号】：TP311.13

【相似文献】