农业文本特征重组向量矩阵算法优化与仿真  

Optimization and Simulation of the Vector Matrix Algorithm for the Reorganization of Agricultural Knowledge Features

作　　者：刘波[1,2] 郭平[1,3] 沈岳 丁德红[1,2] 

机构地区：[1]湖南农业大学信息科学技术学院,湖南长沙410128 [2]湖南省农村农业信息化工程技术研究中心,湖南长沙410128 [3]湖南华南光电科技股份有限公司技术中心,湖南常德415000

出　　处：《软件》2015年第4期6-13,共8页Software

基　　金：国家科技部十二五科技支撑计划课题(2011BAD21B03);国家科技部国家星火计划项目(2011GA770001);国家科技部科技支撑计划课题(2012BAD35B05;2012BAD35B07);湖南省科技厅重点项目(2013FJ1014);2014湖南省教育厅科研一般项目(14C0542);2014湖南省科技厅湘西南农村信息化服务湖南省重点实验室开放基金课题;2013年湖南农业大学团委发明制作科创项目(201401);湖南农业大学2013大学生创新项目(XCX130302014310)资助课题

摘　　要：本文以湖南国家农村农业信息化示范省建设项目中农业知识推送研究成果为基础,提出一个基于用户偏好变化的农业文本特征重组向量矩阵算法。首先,依据用户对农业知识访问的关键词诉求,通过关键词分频策略对推荐信息进行特征提取;其次,基于关键词参照库进行特征重组,得到标准特征词,改进向量空间模型中TF-IDF权重计算方法,设计关键词和特征词的相似度加权值反映用户偏好的变化。通过仿真实验与其它算法对比,此算法在降维的同时,对文本推荐具有较高的准确率和良好的区分度。Based on the agricultural knowledge pushing research results of Hunan National Rural and Agricultural Information Demonstration Projects, a vector matrix algorithm for the reorganization of agricultural knowledge characteristics is put forward in this paper on the base of users' preferences(also for term recombine VSM weighting keywords and term max similarity), Firstly, in accordance with users' demand for the key words of agricultural knowledge, recommended information is feature extracted with the keyword strategy. Secondly, the standard feature words are obtained through keyword reference library features reorganization. Then TF-IDF weight calculation method of the vector space model is improved accordingly. Finally, the keywords and features' similarity weights are designed to show the diversification of users' preferences. Compared with other algorithms through a series of simulation, it proves that the algorithm not only can reduce the dimension, but also is of high accuracy and good discrimination.

关 键 词：矢量矩阵 向量空间模型 推荐算法 特征重组 

分 类 号：TP391.1[自动化与计算机技术—计算机应用技术]