Open wxy015513 opened 1 year ago
Thanks for your question. You can check the optimal value of alpha1 and alpha2 in Fig. 5 of the paper. Overall, DAP-kh achieves the best performance as the impact of high-degree nodes is more significant and can amplify bias. Across the three datasets, we set alpha2 to 0, which corresponds to beta being 0 in the paper. For alpha1, the optimal values are 0.8 for Gowalla, 1.2 for Yelp, and 1.6 for Amazon-Book.
Ok, thanks, and I have some questions about the implementation of cluster P. How do you achieve it?
You can use "from sklearn.cluster import KMeans" to cluster nodes
I know, but i want to know how can i achieve clusters P is searched in {1, 5, 10, 20, 30, · · · , 70}
I manually searched for it. Generally, 10~50 is sufficient.
Can I have one of your contact information?I don't know how to implement it and want to get your instructions.
ok。wechat: zhangyi_meicrs
Can you tell me the optimal value of alpha1 and alpha2?