【摘要】 根据历年的产业网络,提出了一种预测任意两个产业将来发生链接可能性的新方法.利用网络分析方法,以中国的投入产出表为研究对象,构建了基于链路动态变化的产业网络预测模型,对比分析了该模型与相似性链路预测模型的预测精度.研究结果表明,历年产业网络的拓扑结构非常接近.依据正交试验设计,在参数选择合理的前提下,本文提出的预测模型具有理想的预测精度.相比于相似性链路预测算法,基于链路动态变化的产业网络预测算法体现了过往信息对产业网络链路预测的重要性,具有更好的预测结果.更多还原

【Abstract】 Based on the industrial networks over the past years, this paper proposes a new methodology that can predict the likelihood of an existing link between any nodes in the future. With the aid of network analysis,a prediction model of the industrial network is built for Chinese input-output tables. Its prediction accuracy is compared with that of the similarity link prediction model. The results show that the industrial network topology structures are very close over the years. The link prediction model has good prediction accuracy as long as the parameters are selected reasonably according to orthogonal experimental design. Compared with the similarity link prediction algorithm, the proposed methodology reflects the importance of past information to the industrial network link prediction, and has better prediction results.更多还原