Question about DBLP data processing

[zxlearningdeep]

Hello, I have some concerns regarding the data pre-processing in your approach. Specifically, the DBLP dataset consists of three views, but in the load_multi function in utils.py, you only utilize the first two views with higher homophily ratios, while discarding the view with a lower homophily ratio. I attempted to reproduce your method by using all three views, but I was unable to achieve the results reported in the paper.