|
|
|
| An Imputation Technique for Missing Context Data Based on Spatial-temporal and Association Rule Mining |
| Wang Yu-xiang Qiao Xiu-quan Li Xiao-feng Meng Luo-ming |
| State Key Laboratory of Networking and Switching Technology, Beijing University of Posts and Telecommunications, Beijing 100876, China |
|
|
|
|
Abstract The context data missing is an inevitable problem of context information processing mechanism, the imputation technique of missing data also is a research hotspot in data mining. However, the existing imputation technique of missing data is not suitable for the flow data form of context information that does not make full use of data relevance between every collecting sensor. Moreover, that does not take spatial-temporal relationship into account. In order to conquer the shortcomings and deficiencies of the existing imputation technique of missing data, this paper proposes an imputation technique for context data missing based on Spatial-Temporal and Association Rule Mining (STARM) to perform spatiality and time series analysis on sensor data, and generate strong association rules to interpolate missing data. Finally, the simulation experiment verifies the rationality and efficiency of STARM through temperature sensor data acquisition. Experiments show that the algorithm is of high accuracy for the imputation of context data missing, such as Simple Linear Regression (SLR) algorithm and the EM algorithm. In addition, which is smaller time and space overhead and can guarantee Quality of Service (QoS) of real-time applications.
|
|
Received: 12 January 2010
|
|
|
|
Corresponding Authors:
Wang Yu-Xiang
E-mail: quanlin@163169.net
|
|
|
|
|
|
|
2010, Vol. 32 
