search

storyandwine / LAGCN

Code and Datasets for "Predicting Drug-Disease Associations through Layer Attention Graph Convolutional Networks"
51 stars 15 forks source link

About datasets #9

Closed youngbo9i closed 2 years ago

youngbo9i commented 2 years ago

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!

storyandwine commented 2 years ago

interaction_matrix.txt Where 1 for Therapeutic.

youngbo9i commented 2 years ago

Thank you very much！

shahinghasemi commented 2 years ago

Hello @storyandwine, Would you please explain about this data? It consists of {0, 1, -1} what do they mean? what's the relationship between this dataset and the main dataset? are they disjoints?

storyandwine commented 2 years ago

0 for non-associations, 1 for therapeutic, for details see http://ctdbase.org/toolTip.ajx?type=actiontype&ixnType=CHEMICAL-DISEASE&code=t , -1 for harmful, for details see http://ctdbase.org/toolTip.ajx?type=actiontype&ixnType=CHEMICAL-DISEASE&code=m

storyandwine commented 2 years ago

Main dataset 1 = {1,-1} 0 = {0} Therapeutic dataset 1= {1}, 0 = {0,-1} @shahinghasemi

shahinghasemi commented 2 years ago

@storyandwine thanks you for your clarification. How did you split this therapeutic data in your work? I guess for interactions (1s) you probably split them into 5 sets, but I am not sure for 0s. Did you use all the 0s in both training and testing? or just like ones, the zeros are split into 5 sets and 4 of them are used in training and one of them in testing?

storyandwine commented 2 years ago

@storyandwine thanks you for your clarification. How did you split this therapeutic data in your work? I guess for interactions (1s) you probably split them into 5 sets, but I am not sure for 0s. Did you use all the 0s in both training and testing? or just like ones, the zeros are split into 5 sets and 4 of them are used in training and one of them in testing?

It's same to issue #5.