摘要
针对当前数据库信息资源检索结果分类方法准确率和查全率较低的问题,提出基于协同微粒群的关系型数据库信息资源检索结果分类方法,将信息资源检索结果进行标准化,计算标准化后检索结果的协方差,并将其最大值定义为数据库信息资源检索结果预处理目标函数。将方差最大值的求解转换成对样本最大值进行求解,引入拉格朗日系数对目标函数求解,期间融合了数据过滤器,实现检索结果预处理。初始化微粒群,通过骨干微粒群法对微粒位置进行更新。计算微粒适应度值,对微粒群之间的相似性进行判断,将相似的微粒群合并。检测当前运行状态满足终止条件与否和信息资源检索结果流结束与否,对未结束的检索结果数据流类中心变化位置进行预测,一直到不再接收资源检索结果,将最终分类结果输出。实验结果表明,所提方法分类准确性和查全性均较为理想,可行性也高于当前相关方法。
In order to improve the accuracy rate and the recall rate of current method to classify retrieval results of database information resource,this article presents a method to classify retrieval results of information resource in relational database based on collaborative particle swarm. Firstly,the information resource search result was standardized and the covariance of standardized retrieval result was calculated. Secondly,the maximum value was defined as the pre-processing objective function of database information resource retrieval result. Thirdly,the solution of the maximum variance was converted to the solution of the maximum value of the sample. Fourthly,the Lagrangian coefficient was used to solve the objective function. In the meantime,the data filter was integrated and the retrieval result was preprocessed. Fifthly,the particle swarm was initialized and the position of particulate was updated by the barebones particle swarm optimization. Moreover,the fitness value of particulate was calculated and the similarity between particle swarms was judged,and then similar particle swarms were merged. In addition,it was necessary to detect whether the current running state satisfied the termination condition and whether the flow of information resource retrieval results was over. Finally,the central change position of unfinished retrieval result data flow was forecasted until the resource retrieval result was no longer received. Thus,we outputted the final classification result. Simulation results show that the accuracy and recall performance of proposed method are ideal. Meanwhile,the feasibility is higher than the current method.
作者
姚丽华
于广州
YAO Li-hua;YU Guang-zhou(Guangdong Ocean University,Zhanjiang Guangdong 524088,China)
出处
《计算机仿真》
北大核心
2019年第1期445-448,共4页
Computer Simulation
关键词
关系型数据库
信息资源
检索结果
分类
Relational database
Information resource
Retrieval result
Classification
