Enterprise Technology Opportunity Identification-Related Data

Generation of training and testing sets Based on the movability in the network, the link prediction problem of the network is transformed into a binary classification problem, obtaining linked and unlinked edges.

Node2vec is used to obtain network node features, and the edges are represented by the average point vector of the node pairs. Then, the training and testing sets are randomly divided in a 7:3 ratio. 2. Generation of prediction results Using node2vec to obtain the features of unlinked points in the network, calculate the vectors of any two point links, input them into the machine learning model, and output the link probability. 3.