LR-DBN does not merge the same skills cases across users

[sachingrover211]

The param table had 5000 parameters for input of 10000, am I missing some property value. I have gone through the example for multiple sub-skills but I have not found anything. Even for test data provided the same skill were present over and over again with different parameters for students.

[yanboxu]

This would depend on what group of parameters you want to fit: per-student,
per-skill or just per-subskill? Then you would need to switch/modify the
"user" and "skill" columns accordingly in the data files because BNT-SM is
supposed to fit parameters per "skill" column.

For per-skill model that merges the same skill, "user" and "skill" columns
are what the users and skills are.

For per-student model that merges the same student, you need to switch the
two columns, "user" column are the skills and "skill" column are the users.
This is what the example data file do.

For per-subskill model that merges everything, set "skill" column to a
constant name say SKILL for all the rows, and "user" column are what the
users are.

On top of this, the "skill" columns are always sorted in the same order for
training and test data; "user" column need to be sorted as well for
per-skill, and per-subskill models, but not necessarily for per-student
model in order to capture the time dependency between skills, like knowing
how to read HAT should help how to read CAT.

Hope this helps.
Yanbo

On Tue, Oct 27, 2015 at 3:25 PM, Sachin Grover notifications@github.com
wrote:

The param table had 5000 parameters for input of 10000, am I missing some
property value. I have gone through the example for multiple sub-skills but
I have not found anything. Even for test data provided the same skill were
present over and over again with different parameters for students.

—
Reply to this email directly or view it on GitHub
#3.

[sachingrover211]

Okay, that was intentional, I thought that was an error. Cool I will just test that. Thanks.

[sachingrover211]

Is there a way to get the complete inference result for the test data and not just a part of it?