Research of Local Similarity Index Based on OWA Integration Operator in Terrorist Network Link Prediction Method

Based on the known terrorist network structure, mining hidden terrorist networks is a matter of great concern to the security agencies of all countries in the world. With the emergence and development of machine learning methods, the use of various attributes of nodes can be accurately and efficiently predict potential terrorist network links. However, due to the inaccuracy of terrorist network node information, powerful machine learning methods cannot work. Based only on the structure of the network, link prediction methods to find probable links have become a hot issue of concern to experts in related fields in various countries. The link prediction performance of the OWA ensemble operator was compared with the link prediction algorithm of a single similarity index. Experimental results show that the OWA algorithm can effectively reduce the variance of the local information link prediction algorithm and improve the chain stability of link prediction.