Informations about data mining

classic Classic list List threaded Threaded
4 messages Options
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Informations about data mining

Gongus
Good morning, I'm an italian student, so first of all excuse me for my not so perfect english.
I'm ignorant about Weka and I'm struggling to understand how it works, I would appreciate any sort of help.

My problem is:

I have multiples GPS tracking file in .plt (from GeoLife Trajectories). I've converted them into .arff files,  like this:

@RELATION 20090510222007.plt

@ATTRIBUTE latitude NUMERIC
@ATTRIBUTE longitude NUMERIC
@ATTRIBUTE date DATE yyyy-MM-dd HH:mm:ss
@ATTRIBUTE class {STUDENT, WORKER, MOTHER}


@DATA
39.99949,116.327384,2009-05-10 22:20:07, STUDENT
39.999572,116.327393,2009-05-10 22:20:12, STUDENT
39.999648,116.327378,2009-05-10 22:20:17, STUDENT
39.999711,116.327378,2009-05-10 22:20:22, STUDENT
39.999797,116.327279,2009-05-10 22:20:27, STUDENT
39.999942,116.327007,2009-05-10 22:20:32, STUDENT
39.999943,116.326776,2009-05-10 22:20:37, STUDENT
..........

If I understand a bit, Weka can predict the class of others GPS track like:

39.99949,116.327384,2009-05-10 22:20:07, ?

Right? Is this a cluster algorithm?
 
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: Informations about data mining

valerio jus

The algorithm predicts one of the available class labels: STUDENT, WORKER, MOTHER, in your example.

This algorithm is called "Classification" algorithm.

Cheers,
Velerio

On 6 Aug 2017 4:09 a.m., "Gongus" <[hidden email]> wrote:
Good morning, I'm an italian student, so first of all excuse me for my not so
perfect english.
I'm ignorant about Weka and I'm struggling to understand how it works, I
would appreciate any sort of help.

My problem is:

I have multiples GPS tracking file in .plt (from GeoLife Trajectories). I've
converted them into .arff files,  like this:

@RELATION 20090510222007.plt

@ATTRIBUTE latitude NUMERIC
@ATTRIBUTE longitude NUMERIC
@ATTRIBUTE date DATE yyyy-MM-dd HH:mm:ss
@ATTRIBUTE class {STUDENT, WORKER, MOTHER}


@DATA
39.99949,116.327384,2009-05-10 22:20:07, STUDENT
39.999572,116.327393,2009-05-10 22:20:12, STUDENT
39.999648,116.327378,2009-05-10 22:20:17, STUDENT
39.999711,116.327378,2009-05-10 22:20:22, STUDENT
39.999797,116.327279,2009-05-10 22:20:27, STUDENT
39.999942,116.327007,2009-05-10 22:20:32, STUDENT
39.999943,116.326776,2009-05-10 22:20:37, STUDENT
..........

If I understand a bit, Weka can predict the class of others GPS track like:

39.99949,116.327384,2009-05-10 22:20:07, ?

Right? Is this a cluster algorithm?




--
View this message in context: http://weka.8497.n7.nabble.com/Informations-about-data-mining-tp41432.html
Sent from the WEKA mailing list archive at Nabble.com.
_______________________________________________
Wekalist mailing list
Send posts to: [hidden email]
List info and subscription status: https://list.waikato.ac.nz/mailman/listinfo/wekalist
List etiquette: http://www.cs.waikato.ac.nz/~ml/weka/mailinglist_etiquette.html

_______________________________________________
Wekalist mailing list
Send posts to: [hidden email]
List info and subscription status: https://list.waikato.ac.nz/mailman/listinfo/wekalist
List etiquette: http://www.cs.waikato.ac.nz/~ml/weka/mailinglist_etiquette.html
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: Informations about data mining

Gongus
In reply to this post by Gongus
Ok, I've created the "training_set" in this way:

@RELATION training_set

@ATTRIBUTE latitude NUMERIC
@ATTRIBUTE longitude NUMERIC
@ATTRIBUTE date DATE "yyyy-MM-dd HH:mm:ss"
@ATTRIBUTE type_of_person {STUDENT, PROFESSOR, GENERAL_WORKER, MOTHER, UNKNOWN}


@DATA
40.001434,116.312406,"2009-04-03 01:17:07",PROFESSOR
40.001418,116.312536,"2009-04-03 01:17:12",PROFESSOR
40.001398,116.312649,"2009-04-03 01:17:17",PROFESSOR
....
40.006161,116.325652,"2009-04-27 03:59:05",STUDENT
40.006161,116.325652,"2009-04-27 03:59:08",STUDENT
40.006165,116.325654,"2009-04-27 03:59:10",STUDENT
40.006141,116.325652,"2009-04-27 03:59:15",STUDENT
....
39.997664,116.319206,"2009-04-25 05:08:12",GENERAL_WORKER
39.997664,116.319207,"2009-04-25 05:08:17",GENERAL_WORKER
39.997665,116.319209,"2009-04-25 05:08:22",GENERAL_WORKER
39.997665,116.319209,"2009-04-25 05:08:27",GENERAL_WORKER
....

and the "testing_set" in this way:

@RELATION test_set
@ATTRIBUTE latitude NUMERIC
@ATTRIBUTE longitude NUMERIC
@ATTRIBUTE date DATE "yyyy-MM-dd HH:mm:ss"
@ATTRIBUTE type_of_person {STUDENT, PROFESSOR, GENERAL_WORKER, MOTHER, UNKNOWN}


@DATA
39.941191,116.347836,"2009-01-23 04:49:07",?
39.941134,116.347841,"2009-01-23 04:49:10",?
39.941244,116.347845,"2009-01-23 04:49:13",?
39.941233,116.347897,"2009-01-23 04:49:18",?
39.941225,116.347932,"2009-01-23 04:49:23",?
39.941113,116.347945,"2009-01-23 04:49:26",?
39.940929,116.34794,"2009-01-23 04:49:58",?
....
I've loaded the training_set on the "Preprocess table" (I didn't touch any other fields) like this:


Then I've loaded the test_set in the "Classify" tab like this:


Why I'm obtaining results like this?


Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: Informations about data mining

valerio jus


> Why I'm obtaining results like this?
> <http://weka.8497.n7.nabble.com/file/n41444/Immagine_3.png>

The result is not wrong. You had such result because you've used the "?" for the class label. This indicates that the class label is missing.

Cheers,
Valerio

>
>
>
>
>
> --
> View this message in context: http://weka.8497.n7.nabble.com/Informations-about-data-mining-tp41432p41444.html
> Sent from the WEKA mailing list archive at Nabble.com.
> _______________________________________________
> Wekalist mailing list
> Send posts to:[hidden email]
> List info and subscription status: https://list.waikato.ac.nz/mailman/listinfo/wekalist
> List etiquette: http://www.cs.waikato.ac.nz/~ml/weka/mailinglist_etiquette.html


_______________________________________________
Wekalist mailing list
Send posts to: [hidden email]
List info and subscription status: https://list.waikato.ac.nz/mailman/listinfo/wekalist
List etiquette: http://www.cs.waikato.ac.nz/~ml/weka/mailinglist_etiquette.html
Loading...