weights of the data sample

[Yurikusu]

I saw the dataset of tgbl-review and I saw the four columns : ts,source,target,weight

How was the weight of the edge calculated? Out of what information exaclty?

I plan to use the model on my own data for Link prediction and I am not quite sure if I can use multiple columns.
I have properties for each node. In my case I try to model something with customer and product nodes. I want to add properties like age and gender to a customer node. Do I have to put all these into one column "weight" ?

Best regards

[shenyangHuang]

Hi Yurikusu,

Thank you for your interest in our work.
The weight for this dataset is the user review score corresponding with the edge, i.e. which score the user gave to this product. You can use multiple columns with your data file in csv format. You would need to write your own data pre-processing code. For more details, please see issue #84

Best,
Andy