PCRW evaluates the path according to the neighborhood of the node, and considering the influence of node type selection at each step on the degree of relevance, a probability selection model is added to each step of the walk, which contributes to the final result more in line with the actual situation. <br><br>However, the above two similarity measurement methods lack consideration of node type label selection, and the meta-path must be given in advance.
正在翻译中..