基于双通道特征融合的WPOS-GRU专利分类方法.PDFVIP

37 3 Vol. 37 No. 3 第 卷第 期 计算机应用研究 录用定稿 Application Research of Computers Accepted Paper 基于双通道特征融合的WPOS-GRU 专利分类方法* a, b a 余本功 ，张培行 (合肥工业大学 a. 管理学院; b. 过程优化与智能决策教育部重点实验室, 合肥 230009) 摘 要：为提高专利文本的自动分类的效率和准确度，提出一种双通道特征融合WPOS-GRU(word2vec and part of speech gated recurrent unit)专利文本自动分类方法。首先获取专利摘要文本，并进行清洗和预处理；然后对专利文本 进行词向量表示和词性标注，并将专利文本分别映射为word2vec 词向量序列和POS 词性序列；最后使用两种特征 通道训练 WPOS-GRU 模型，并对模型效果进行实验分析。通过对比传统专利分类方法和单通道专利分类方法，双 通道特征融合的 WPOS-GRU 专利分类方法提高了分类效果。本文提出的方法节省了大量的人力成本，提高了专利 文本分类的准确度，更能满足大量专利文本分类任务自动化高效率的需要。 关键词：专利分类；词性标注；特征融合；门限递归单元 中图分类号：TP 391 doi: 10.3969/j.issn.1001-3695.2018.08.0628 WPOS-GRU patent classification method based on dual channel feature fusion Yu Bengonga,b, Zhang Peihanga (1. School of Management , b. Key Laboratory of Process Optimization Intelligent Decision-Making of Ministry of Education , Hefei University of Technology, Hefei 230009, China) Abstract: In order to improve the efficiency and accuracy of patent text automatic classification, this paper proposes a two-channel feature fusion WPOS-GRU patent text automatic classification method . Firstly, the patent summary text is obtained, cleaned and pretreated, then the patent text is represented by word vector and part-of-speech tagging, and the patent text is mapped into word 2vec word vector sequence and POS part-of-speech sequence, respectively. Finally, WPOS-GRU model is trained by tw