社会网络二元不对称关系预测中的非独立性检验问题

The Problem of Non-Independence Testing During Predicting Binary Dis-Symmetry Relation in the Social Network

社会网络中普遍存在着二元不对称关系,在网络动态演化过程中对二元不对称关系预测是基于现有的大量网络结构信息,判定信息的有效性是提高预测效率和预测结果正确率的重要保证.文章针对二元不对称关系预测提出了变量非独立性假设以及检验算法,并通过建立正态性检验算法进一步提高非独立性检验的效率.实例验证表明本文所提出的方法是合理的.

There is a common binary dis-symmetry relation in the network, and the prediction of the binary dis-symmetry relation is based on the existing network information in the process of network dynamic evolution, and the validity of deci- sion information is an important guarantee to improve the efficiency and accuracy of prediction. In this paper, the non-independence hypothesis and testing algorithm are put forward, and the normality testing algorithm is established to further improve the efficiency of non independence testing. Finally, the example shows that the method proposed in this paper is reasonable.