Classify time series data

classic Classic list List threaded Threaded
5 messages Options
Reply | Threaded
Open this post in threaded view
|

Classify time series data

Theo
Hi. I am trying to classify students' performance (pass or fail) utilizing their online traces when accessing a webpage of an online course. My dataset has 100 rows (one per student) and 11 columns. 10 columns for online behavior attributes (e.g. TotalLogins, TotalFilesAccessed, TotalForumPosts etc) and 1 (the last) column for the class (pass or fail). StudentID,TotalLogins,TotalFilesAccessed,TotalForumPosts,...,Class 1,45,8,5,..., pass 2,53,12,3,...,fail .... I can run and compare the classic classification algorithms on my dataset using the WEKA GUI. I need to target my study towards time series classification, since the attribute values are accumulated over time. 1)How to shape a new dataset in order to contain the weekly values of the same attributes? For example: StudentID, LoginsWeek1, LoginsWeek2, ...,LoginsWeekN, FilesAccessedWeek1, FilesAccessedWeek2, ..., FilesAccessedWeekN, ..., Class 2)Which steps to follow and which algorithm to use in the GUI to take advantage of the temporal structure of the new dataset? Thanks. T

Sent from the WEKA mailing list archive at Nabble.com.

_______________________________________________
Wekalist mailing list -- [hidden email]
Send posts to [hidden email]
To unsubscribe send an email to [hidden email]
To subscribe, unsubscribe, etc., visit https://list.waikato.ac.nz/postorius/lists/wekalist.list.waikato.ac.nz
List etiquette: http://www.cs.waikato.ac.nz/~ml/weka/mailinglist_etiquette.html
Reply | Threaded
Open this post in threaded view
|

Re: Classify time series data

Eibe Frank-3

On Mon, Apr 5, 2021 at 1:44 PM Theo <[hidden email]> wrote:
Hi. I am trying to classify students' performance (pass or fail) utilizing their online traces when accessing a webpage of an online course. My dataset has 100 rows (one per student) and 11 columns. 10 columns for online behavior attributes (e.g. TotalLogins, TotalFilesAccessed, TotalForumPosts etc) and 1 (the last) column for the class (pass or fail). StudentID,TotalLogins,TotalFilesAccessed,TotalForumPosts,...,Class 1,45,8,5,..., pass 2,53,12,3,...,fail .... I can run and compare the classic classification algorithms on my dataset using the WEKA GUI. I need to target my study towards time series classification, since the attribute values are accumulated over time. 1)How to shape a new dataset in order to contain the weekly values of the same attributes? For example: StudentID, LoginsWeek1, LoginsWeek2, ...,LoginsWeekN, FilesAccessedWeek1, FilesAccessedWeek2, ..., FilesAccessedWeekN, ..., Class 2)Which steps to follow and which algorithm to use in the GUI to take advantage of the temporal structure of the new dataset? Thanks. T

Sent from the WEKA mailing list archive at Nabble.com.
_______________________________________________
Wekalist mailing list -- [hidden email]
Send posts to [hidden email]
To unsubscribe send an email to [hidden email]
To subscribe, unsubscribe, etc., visit https://list.waikato.ac.nz/postorius/lists/wekalist.list.waikato.ac.nz
List etiquette: http://www.cs.waikato.ac.nz/~ml/weka/mailinglist_etiquette.html

_______________________________________________
Wekalist mailing list -- [hidden email]
Send posts to [hidden email]
To unsubscribe send an email to [hidden email]
To subscribe, unsubscribe, etc., visit https://list.waikato.ac.nz/postorius/lists/wekalist.list.waikato.ac.nz
List etiquette: http://www.cs.waikato.ac.nz/~ml/weka/mailinglist_etiquette.html
Reply | Threaded
Open this post in threaded view
|

Re: Classify time series data

Theo
Hi Eibe.
Thanks for your reply.
And please accept my apology for my first post's format. Due to a mistype,
all paragraph marks disappeared. So, everything was posted as a unique
paragraph.

I had already located and read the page you suggested. Unfortunately, I
couldn't apply the advises in that post (and in the posts it contained) to
my case.

So, I would appreciate any further suggestions.
For example, any web resource that it could guide me how to appropriately
shape my dataset, would be a good starting point.

Regards,
Theo



--
Sent from: https://weka.8497.n7.nabble.com/
_______________________________________________
Wekalist mailing list -- [hidden email]
Send posts to [hidden email]
To unsubscribe send an email to [hidden email]
To subscribe, unsubscribe, etc., visit https://list.waikato.ac.nz/postorius/lists/wekalist.list.waikato.ac.nz
List etiquette: http://www.cs.waikato.ac.nz/~ml/weka/mailinglist_etiquette.html
Reply | Threaded
Open this post in threaded view
|

Re: Classify time series data

Theo
Hi again.
It seems that there are not any other replies for my post.
Should I conclude that there are no other resources / documentation than
those suggested by Eibe?
Theo



--
Sent from: https://weka.8497.n7.nabble.com/
_______________________________________________
Wekalist mailing list -- [hidden email]
Send posts to [hidden email]
To unsubscribe send an email to [hidden email]
To subscribe, unsubscribe, etc., visit https://list.waikato.ac.nz/postorius/lists/wekalist.list.waikato.ac.nz
List etiquette: http://www.cs.waikato.ac.nz/~ml/weka/mailinglist_etiquette.html
Reply | Threaded
Open this post in threaded view
|

Re: Classify time series data

Eibe Frank-2
Administrator
I had totally forgotten about the RnnSequenceClassifier in WekaDeeplearning4j!

Cheers,
Eibe

> On 11/04/2021, at 7:57 PM, Theo <[hidden email]> wrote:
>
> Hi again.
> It seems that there are not any other replies for my post.
> Should I conclude that there are no other resources / documentation than
> those suggested by Eibe?
> Theo
>
>
>
> --
> Sent from: https://weka.8497.n7.nabble.com/
> _______________________________________________
> Wekalist mailing list -- [hidden email]
> Send posts to [hidden email]
> To unsubscribe send an email to [hidden email]
> To subscribe, unsubscribe, etc., visit https://list.waikato.ac.nz/postorius/lists/wekalist.list.waikato.ac.nz
> List etiquette: http://www.cs.waikato.ac.nz/~ml/weka/mailinglist_etiquette.html
_______________________________________________
Wekalist mailing list -- [hidden email]
Send posts to [hidden email]
To unsubscribe send an email to [hidden email]
To subscribe, unsubscribe, etc., visit https://list.waikato.ac.nz/postorius/lists/wekalist.list.waikato.ac.nz
List etiquette: http://www.cs.waikato.ac.nz/~ml/weka/mailinglist_etiquette.html