학술논문
用户行为数据分析下的信息推送系统的设计 / Design of information push system under user behavior data analysis
Document Type
Academic Journal
Author
Source
现代电子技术 / Modern Electronics Technique. 40(17):158-166
Subject
Language
Chinese
ISSN
1004-373X
Abstract
针对在信息推送过程中难以计算出用户行为数据分析下信息推送指标的权重输出,存在信息推送针对性差,信息推广转化率低的问题,提出一种基于模糊遗传的用户行为数据分析下的信息推送系统设计方法.将互联网用户抽象为结构化数据信息本体,计算出用户行为数据下各词汇的TF-IDF指标,建立用户行为数据下的行为-主题分布模型和主题-词汇分布模型,得到用户行为数据下所有主题-词汇的联合分布,利用向量模型匹配关键词思想理论确定关键词在文档中的权重,通过用户查询与结果记录词条间的关联性计算,得到用户行为数据下不同信息推送指标的权重输出,由此完成用户行为数据分析下的信息推送系统设计.实验结果表明该方法能够有效提高信息推送的准确性,且可扩展性较强.
The current method to push information is difficult to calculate the output weight of information push index un-der user behavior data analysis,and has poor pertinency and low percent conversion of information push. Therefore,a fuzzy ge-netic based design method of information push system under user behavior data analysis is proposed. The Internet users are ab-stracted as structural data information ontology by means of the method to calculate the TF-IDF index of each vocabulary under user behavior data. The behavior-theme distribution model and theme-vocabulary distribution model under user behavior data were established to get all theme-vocabulary joint distributions under user behavior data. The theory of vector model marching keyword is used to determine the weight of keyword in the file. The correlation between entries is calculated by means of user query and result recording to get the output weight of different information push indexes under user behavior data,so as to de-sign the information push system under user behavior data analysis. Experimental results show that the method can improve the accuracy of information push effectively,and has strong scalability.
The current method to push information is difficult to calculate the output weight of information push index un-der user behavior data analysis,and has poor pertinency and low percent conversion of information push. Therefore,a fuzzy ge-netic based design method of information push system under user behavior data analysis is proposed. The Internet users are ab-stracted as structural data information ontology by means of the method to calculate the TF-IDF index of each vocabulary under user behavior data. The behavior-theme distribution model and theme-vocabulary distribution model under user behavior data were established to get all theme-vocabulary joint distributions under user behavior data. The theory of vector model marching keyword is used to determine the weight of keyword in the file. The correlation between entries is calculated by means of user query and result recording to get the output weight of different information push indexes under user behavior data,so as to de-sign the information push system under user behavior data analysis. Experimental results show that the method can improve the accuracy of information push effectively,and has strong scalability.