When I use your original dataset and the lrssl dataset, the resul is normal.But when I run the program on Fdataset and Cdataset, the AUPR is 0.0005 and the AUC is about 0.5. It's abnormal.

layers.py文件中第22行

x = tf.nn.dropout(x, 1-self.dropout)

tf.nn.dropout的第2个参数含义为每个元素的值设置为0的概率，所以代码是否应该写为

x = tf.nn.dropout(x, self.dropout)

Hello, I see that the Therapeutic dataset is also used in your work, but it is not provided in the project. Is it convenient to provide it, thank you!

Hello, thanks for your great work, is that possible please send me all the raw data (hasn't been processed by your method) to me? (my email: [email protected])
another question: is that possible to add some further diseases information e.g. side effects, targets for drugs etc. to your framework?

thanks in advance.

Your drug similarity matrix does not match the Jaccard similarity matrix properties, for example, the diagonal elements should all be 1 but you have them all at 0. Am I missing something? Thank you in advance.

在main.py文件中的第16行

adj = constructHNet(train_drug_dis_matrix, drug_matrix, dis_matrix)

utils.py文件中

def constructHNet(drug_dis_matrix, drug_matrix, dis_matrix):
    mat1 = np.hstack((drug_matrix, drug_dis_matrix))
    mat2 = np.hstack((drug_dis_matrix.T, dis_matrix))
    return np.vstack((mat1, mat2))

此处构造了异构网络，后续将作为input graph G，但是好像并没有如论文中所言（论文中的构造异构网络部分）对相似性矩阵进行归一化操作？

您好！
我按照说明运行了main.py 文件 ，但运行结束没有得到任何结果性的文件如 一个新的drug-dis矩阵，无法判断预测的具体情况！

hello,
I need to drug target feature and the MESH of disease ( the raw data that drug similarity and disease similarity are extracted from these )
thank you for giving it to me ..

bonjour,
s'il vous plait de me donner l'article scientifique liée a ce code pour bien comprendre et merci d'avance
cordialement.

Hi, your repository gives drug similarity based on the Jaccard algorithm and disease-disease similarity based on MeSH is it? Thank you in advance

Hello, please consider this code below which I copied from the repository:

def cv_model_evaluate(interaction_matrix, predict_matrix, train_matrix):
    test_index = np.where(train_matrix == 0)
    real_score = interaction_matrix[test_index]
    predict_score = predict_matrix[test_index]
    return get_metrics(real_score, predict_score)

As I understood you're only considering the 0s(real zeros and the ones we set to zeros). Here's my question: Why didn't we consider 1s? this way we're missing the 1s that were really 1 and 0s that has been predicted 1s which IMO are considerable for our model performance evaluation.

Thanks.

Could you please provide the name of corresponding drugs and diseases in the data files? Thank you!

I was hoping you could send me a copy of that dataset, it would be of great help in our research. The Agreement is attached.
Thanks in advance!
email:[email protected]

I'd like to compare my method with yours, but it feels kind of unfair since my model also uses target-related information (drug-target interactions, disease-target interactions, protein-protein interactions). It would be really helpful if you can show me how to use LAGCN to handle tri-layer network.

utils.py文件中第18行开始

def dropout_sparse(x, keep_prob, num_nonzero_elems):
    noise_shape = [num_nonzero_elems]
    random_tensor = keep_prob
    random_tensor += tf.random_uniform(noise_shape)
    dropout_mask = tf.cast(tf.floor(random_tensor), dtype=tf.bool)
    pre_out = tf.sparse_retain(x, dropout_mask)
    return pre_out*(1./keep_prob)

该函数的含义好像是：每个元素以keep_prob的概率保留原值并*(1/keep_prob)，并且以1-keep_prob的概率将值设置为0，
而代码的意思是随机的将一些值设置为0，好像并没有体现以1-keep_prob将值设为0的意思？

hi
Can you give me a source for data and The result obtained in this program For comparison with my result please
[email protected] my email
cordially.

Hello, do you have the names of the drugs and diseases corresponding to this dataset? thanks.

您好！
请问您在基础数据集上与其它六种方法比较的程序是发表在哪利呀，或者是从哪里找到的呀