摘要
针对军事领域用户个性化信息需求分析缺乏有效工程化技术支撑问题,笔者提出了一种通过数据驱动的用户军事信息需求分析挖掘方法。该方法借助门户系统采集用户相关的文本关联数据,然后学习回归模型对关联数据进行筛选过滤,最后建立需求概率模型对需求相关度进行排序。该方法支持开放的需求词典,同时能够有效应对噪声数据干扰。搭建的演示验证系统实验结果证明了提出方法的实用性和有效性。
There is a lack of effective engineering technical support for user personalized information demand analysis in military field,so the author presents a data-driven user's military information demand analysis mining method.This method uses the portal system to collect user-related text-related data,then the regression model is used to filter the association data,finally,the demand probability model is used to sort the demand correlation.The method supports an open demand dictionary,while being able to effectively cope with noise data interference.The experimental results show that the proposed method is practical and effective.
作者
戴礼灿
Dai Lican(Southwest China Institute of Electronic Technology,Chengdu Sichuan 610036,China)
出处
《信息与电脑》
2017年第10期172-175,共4页
Information & Computer
关键词
需求挖掘
门户交互
需求概率模型
相关度排序
demand mining
portal interaction
demand probability model
relevance ranking