How to compute the "trip" argument?

[LeoGautheron]

In your files "train_FLRML.m" and "train_MFLRML.m" you are calling your proposed method with an argument called "trip".

I assume this has to be computed using the label information about the dataset.
Could you provide the code that constructs such matrix?

[highan911]

The matrix "trip" is in the shape t*3, in which each row contains a triple of instance ids, which can be named as:
[ my_id, positive_id, negative_id ]

The method to generate the triples from class labels is simple, such as:
my_id = a random id
positive_id = another random id with the same class label
negative_id = another random id with a different class label

Or maybe there are some other data sets without "one class label" for each sample, the triples can be generated differently in other ways.