Role of degree and weighted coreness based on endpoints in link prediction

Many researchers propose link prediction models based on node similarity. Among all models, found that the endpoint influence plays an important role in evaluating similarity between endpoints. For influence, we consider possessing a large and extensive maximum connected subgraph can strongly attract other nodes. After thorough research, coreness describe aggregation degree of neighbors may be used to largest endpoint. In order create model, repeat our experiments eight real benchmark datasets after combining weighted coreness. The experimental results illustrate positive synthetical for measuring prediction.