Abstract: In order to improve the accuracy of the estimated missing data in Wireless Sensor Network (WSN), a self-decision interpolation algorithm was proposed. The algorithm selected different estimation strategies of missing data according to the spatial correlation of the data sets and the continuity of missing data, then introduced the Auto-Regressive and Moving Average (ARMA) model into the study of missing data interpolation. In corresponding to the traditional missing value estimation algorithm, the proposed algorithm not only considered the characteristics of wireless sensor networks, but also took the characteristics of the data themselves into account. The experimental results on the real data sets show that the proposed algorithm improves the precision of the estimation for missing data.

Key words: Wireless Sensor Network (WSN), missing data, interpolation algorithm, Auto-Regressive and Moving Average (ARMA) model, spatial correlation

摘要： 为了提高无线传感器网络(WSN)中缺失数据估计值的精度，提出了一种自决策插值算法。该算法能够根据数据集的空间相关性以及缺失数据的连续性选择不同的缺失数据估计策略，并将自回归滑动平均(ARMA)模型引入到对缺失数据插值的研究中。与传统缺失值估计算法相比，该算法不仅考虑到无线传感器网络的特性，而且考虑到数据集本身的特性。在真实数据集上测试结果表明，该算法提高了对缺失值估计的精度。

关键词: 无线传感器网络, 缺失数据, 插值算法, 自回归滑动平均模型, 空间相关性