摘要
目前的数据库信息检索方法检索耗时长、精度低,提出基于粒子群的分布式文献数据库需求信息自适应检索方法。利用局部过滤器对单个阅读器收集的数据进行处理,采用时间延迟将数据依照时间戳进行排序,依据排序结果将过期数据删除。检测数据库信息,判断标签数据信息是否存在冲突,通过全局过滤器将存在冲突的冗余数据删除。设置粒子群运行的初始参数,并找到当前最佳粒子位置。判断当前粒子位置是否为最佳位置,如果不是,那么利用粒子运动学方程和运动规则实现粒子位置更新。计算粒子群寻优的适应度函数,依据适应度函数和迭代终止条件,找到最佳粒子位置,即最优信息检索结果。实验结果表明,上述方法信息检索结果精度和检索效率高,用户满意度可达99%。与当前相关方法相比,所提方法具有明显优势,具有可行性。
This paper proposes an adaptive retrieval method for requirement information in distributed document database based on particle swarm. Firstly, we used local filter to process data collected by single reader and used time delay to sort data based on time stamp, and then deleted stale data based on the sorting result. Moreover, we detected database information and judged whether there was a conflict in tag data information. Meanwhile, we deleted the conflicting redundant data through global filter. In addition, we set initial parameters for particle swarm operation and found the best particle position to judge whether current particle position was the best. If not, we needed to use particle kinematics equation and motion rule to achieve the update of particle position. Finally, we calculated the fitness function of particle swarm optimization. Based on the fitness function and the iterative termination condition, we could find the best particle position, namely optimal information retrieval result. Simulation results show that the proposed method has high accuracy of information retrieval and high retrieval efficiency. Meanwhile, the degree of satisfaction can reach 99%. Compared with current methods, the proposed method has obvious advantage and feasibility.
作者
张祥合
ZHANG Xiang-he(Department of Engineering and Technology Edition Journal,Jilin University,Changchun Jilin 130022,China)
出处
《计算机仿真》
北大核心
2018年第9期409-412,共4页
Computer Simulation
关键词
分布式文献数据库
信息
检索
Distributed document database
Information
Retrieval
