Share your experience about Moodle Analytics models

Share your experience about Moodle Analytics models

by Motasem Al Haj Ali -
Number of replies: 5

Hi all, 


It would be great if you can share with us your experience using the Moodle Analytics or enabling Moodle prediction models. 

Such as what type of problems you had by enabling them?, what do you think about the practicality of using them or the accuracy of their prediction? any problems with the server

Thanks in advance

Mo

Average of ratings: Useful (1)
In reply to Motasem Al Haj Ali

Re: Share your experience about Moodle Analytics models

by Leonardo Garcia Razera -
I've just started using the models and landed here googling for some source models to import.

The accuracy of the predictions comes along with de quality of the data. What i did is to ensure the models are tracking courses from the categories i want them to receive information and retrieve predictions. Lets see the results next week.
In reply to Motasem Al Haj Ali

Re: Share your experience about Moodle Analytics models

by Shufeng Bai -
Dear Mo,
I am starting to use it. But how do I know what criteria is used for predicting "upcoming activities due"? I have googled around but could not find a good answer.
In reply to Shufeng Bai

Re: Share your experience about Moodle Analytics models

by John Provasnik -
Picture of Particularly helpful Moodlers Picture of Testers
blue automatically appears for all activities that do not have a due date, or have a future 'expected completed on' date.
In reply to Motasem Al Haj Ali

Re: Share your experience about Moodle Analytics models

by Robert Keys -
We have just enabled Analytics and are trying to wrap our heads around it. We have the Students at Risk of dropping out model running at Seven Days and One Month after course start. In the one course I'm looking at I see a student flagged in the Seven Days insight report but not in the One month and vice versa two students in the One Month report but not the Seven day. Looking at their course activity I would expected all of them to be in both reports? I can see an improperly flagged student can be marked as Not Applicable or Incorrectly Flagged but how can I indicate a student has been missed?

Also in the select other insights... drop down menu but versions are labeled as "Student at risk of dropping out" rather than the model names. Is this normal?

Thanks,
Rob
In reply to Robert Keys

Re: Share your experience about Moodle Analytics models

by Dan Marsden -
Picture of Core developers Picture of Particularly helpful Moodlers Picture of Peer reviewers Picture of Plugin developers Picture of Plugins guardians Picture of Testers Picture of Translators
cool! - your 7 day and 1 month models will include quite different training data-sets and it's quite hard to identify why this might have occurred without digging (very deep) into the data generated by your models - also - are you using the more accurate tensorflow backend or the built-in php models?

We've found it important to use the predictions as one component of the bigger picture, rather than relying on them on their own too.