东北大学学报:自然科学版 ›› 2014, Vol. 35 ›› Issue (3): 337-341.DOI: 10.12068/j.issn.1005-3026.2014.03.008

• 信息与控制 • 上一篇    下一篇

基于协作过滤的传感器数据补全方法

李飞,张斌,高岩,张鑫龙   

  1. (东北大学 信息科学与工程学院, 辽宁 沈阳110819)
  • 收稿日期:2013-06-17 修回日期:2013-06-17 出版日期:2014-03-15 发布日期:2013-11-22
  • 通讯作者: 李飞
  • 作者简介:李飞(1975-),男,辽宁辽阳人,东北大学秦皇岛分校讲师,博士;张斌(1964-),男,辽宁本溪人,东北大学教授,博士生导师.
  • 基金资助:
    国家自然科学基金资助项目(61073062);辽宁省自然科学基金资助项目(20102061);沈阳市科技基金资助项目(F12-277-1-80,F12-029-2-00,F11-264-1-35);中央高校基本科研业务费专项资金资助项目(N110604002,N120604003).

Sensor Data Complement Method Based on Collaborative Filtering

LI Fei, ZHANG Bin, GAO Yan, ZHANG Xinlong   

  1. School of Information Science & Engineering, Northeastern University, Shenyang 110819, China.
  • Received:2013-06-17 Revised:2013-06-17 Online:2014-03-15 Published:2013-11-22
  • Contact: ZHANG Bin
  • About author:-
  • Supported by:
    -

摘要: 如何对缺失的传感器监测数据进行补全是物联网信息感知过程中的一个关键问题.针对这一问题,提出了一种基于协作过滤的传感器数据补全方法.该方法利用传感器之间的时空相关特性,考虑到同一区域的传感器或同一传感器的不同监测周期中相应的监测数据具有很大的相似性这一特点,通过对缺失数据的传感器进行分类,分别使用不同的相似评价方法选取相似传感器,以保证估计值的准确性.结果表明,该方法对环境变化幅度较大时段的缺失数据进行估值的效果要优于其他方法.

关键词: 物联网, 传感器, 数据缺失, 数据补全, 协作过滤

Abstract: Completing the lack data monitored by the sensors is a key problem of the information sensing process in the Internet of things. In order to solve this problem, a method was proposed, which can complete sensor data by using collaborative filtering. Considering that there are a lot of similarities among monitor data from sensors in the same area or from the one with different periods, properties of spacetime correlation between sensors were adopted in the proposed method. Different similarity evaluation was used to select similar sensors by classifying sensors with missing data in order to ensure the accuracy of the estimate. The results showed that using this method to estimate the missing data was better than other methods when there are large changes in the environment.

Key words: Internet of things, sensor, missing data, data completion, collaborative filtering

中图分类号: