Divided pseudo session trajectory

[Kyraaaaaaaa]

Hi,
When I checked the code for generating a pseudo session trajectory, I found the logic of the code was wrong. As you mentioned in the paper, the work should "Divide user check-in records into several trajectories with 24-hour intervals, excluding trajectories that contain only one check-in record. ". But for the code, it divided the trajectories by the time differences between two records near each other instead of the time difference between the current record and the start record of the trajectory.
Can you check for this? Thank you !

[Kyraaaaaaaa]

And I found that for the current division code, the longest trajectory has 238 check-in records.

[neolifer]

The logic is correct. The 24-hour interval is meant to be a sliding window. I also found it quite ambiguous when I first read that from other papers. Here I just followed what they write for data preprocessing. The reason people do this I guess is that the data is too sparse.