摘要
作者重名消歧是一个重要又复杂的研究课题,在科技文献检索工作中,作者重名问题势必会降低文献检索的效率和准确性,影响工作进度。提出一种改进粒子群算法优化的BP(Back Propagation)神经网络算法,以解决作者重名消歧问题。首先引入Beta分布的动态惯性权重,提高算法全局搜索能力;其次利用改进粒子群算法优化的权值和阈值,作为BP神经网络的初始权值和阈值进行模型训练,以加快模型训练速度;最后通过特征评价函数过滤式选取排序较优的M维特征子集作为输入层特征向量训练模型,得到最终预测结果,从而精确区分重名的作者。实验研究表明,该模型对重名作者身份的预测准确率可达89.01%,证明了该算法的有效性。
The author’s name and disambiguation is an important and complicated research topic.In the retrieval of scientific literature,the author’s name problem will inevitably reduce the efficiency and accuracy of literature retrieval and affect the progress of the work.In this paper,a back propagation(BP)neural network algorithm with improved particle swarm optimization is proposed to solve the problem of author’s name disambiguation.Firstly,the dynamic inertia weight of Beta distribution is introduced to improve the global search ability of the algorithm.Secondly,the weight and threshold of the improved particle swarm optimization algorithm are used as the initial weight and threshold of BP neural network to train the model to speed up the training of the model.The feature evaluation function is used to filter and select the M-dimensional feature subsets with better ranking as the input layer feature vector training model to obtain the final prediction result,so as to accurately distinguish the authors of the duplicate names.The experimental results show that the prediction accuracy of the model can be improved to 89.01%,which proves the effectiveness of the algorithm.
作者
仇国华
赵华
QIU Guo-hua;ZHAO Hua(College of Computer Science and Engineering,Shandong University of Science and Technology,Qingdao 266590,China)
出处
《软件导刊》
2020年第3期111-115,共5页
Software Guide
基金
教育部人文社会科学研究青年基金项目(16YJCZH154)。
关键词
重名消歧
PSO算法
BP神经网络
动态惯性权重
特征评价函数
duplicate disambiguation
PSO algorithm
BP neural network
dynamic inertia weight
feature evaluation function
