半监督的网络科技信息分类模型

发布时间：2018-06-10 16:23

本文选题：网络科技信息 + 网络科技信息分类模型　；参考：《现代图书情报技术》2014年11期

【摘要】：【目的】开放的网络科技信息网页内容之间区分度较小,传统基于规则和统计学习的方法无法满足网络科技信息网页分类的具体应用需求。【方法】通过深入分析网络科技信息主题网页的内容和结构,利用开放本体等资源实现领域特征的学习,构建半监督的网络科技信息分类模型。【结果】实验结果表明提出的方法在网络科技信息分类实验中的精度、召回率和F1值分别达到0.9016、0.8756和0.8884,相比贝叶斯方法具有明显优势。【局限】该方法在应用到其他类别的网络科技信息分类时,仍然需要领域专家提供相关领域的核心种子特征。【结论】该方法可以满足网络科技信息深度加工的需求,实现有效的网络科技信息网页分类。
[Abstract]:[objective] Open web science and technology information web pages have a relatively small degree of differentiation, The traditional rule-based and statistical learning methods can not meet the specific application needs of the classification of web pages of network science and technology information. [methods] through in-depth analysis of the content and structure of the subject pages of network science and technology information, Using open ontology and other resources to realize domain feature learning, a semi-supervised classification model of network science and technology information is constructed. [results] the experimental results show the accuracy of the proposed method in the network science and technology information classification experiment. The recall rate and F1 value are 0.9016, 0.8756 and 0.8884, respectively, which have obvious advantages over Bayes method. [limitation] this method is applied to other types of network scientific and technological information classification. It is still necessary for domain experts to provide the core seed features of related fields. [conclusion] this method can meet the needs of the advanced processing of network science and technology information and realize the effective classification of web pages of network science and technology information.
【作者单位】：中国科学院文献情报中心;
【基金】：中国科学院文献情报能力建设专项“网络科技信息自动监测系统二期建设”项目(项目编号:院1306) 国家“十二五”科技支撑计划课题“科技知识组织体系共享服务平台建设”(项目编号:2011BAH10B03)的研究成果之一
【分类号】：TP393.0

【参考文献】