原创科技复杂网络的反演模型研究

Inversion model of an original science complex network

为减少先进科技跟踪成本,研究了原创科技复杂网络的反演模型,将传统的科技信息直接获取变为间接推理。该模型首先以约束反演方法来提取复杂网络中各个节点的社会属性和技术属性,并对其进行量化;然后研究属性间关联关系的计算方法,形成统一的分值矩阵模型,并量化复杂网络模型的各个节点间的权值;最后对模型进行分析形成多种优化目标的计算方法,并采用约束反演推理策略进行复杂网络最优路径求解,其中引入变量事件分类以加快反演速度。实验结果表明,该模型能够快速反演出先进科技的来源,并具有良好的可信度。

In an attempt to reduce tracking costs for advanced science, an inversion model of an original science complex network has been developed in which indirect reasoning is used rather than obtaining information directly. The model first extracts social and technical attributes of each node, sampled according to the weight of each node. Secondly the method of calculating the relationship between those properties is provided. Applying a unified score matrix model allows the complex weight of each edge of the network model to be calculated. Finally several optimization methods are used. We can combine the constrained inversion process with several reasoning approaches to establish an optimal path for the complex network. Moreover, our method speeds up the inversion by applying a variable event classification. The experimental results show that the model can find the sources of advanced technology in reasonable time and also provide sufficient accuracy.