Query about classification of data in WEKA

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

Query about classification of data in WEKA

Juan Sebastian Mejia
Hello everyone, I'm new using WEKA. I am developing a solar radiation prediction project with some meteorological variables as input of the model: Hours of solar brightness, cloudiness, maximum temperature, minimum temperature, relative humidity, etc.

I used WEKA as a tool to identify which of these input variables are the most relevant and correlated in relation to solar radiation.

I chose Select Attributes: PrincipalComponents as an attribute evaluator and the Ranker search method.

I got the following results:

Evaluator:    weka.attributeSelection.PrincipalComponents -R 0.95 -A 5
Search:       weka.attributeSelection.Ranker -T -1.7976931348623157E308 -N -1
Relation:     solarradiation
Instances:    1275
Attributes:   8
              HorasBrillo
              Nubosidad
              TempMax
              TempMin
              HumedadRelativa
              VeloViento
              Precipitacion
              Radiacion
Evaluation mode:    evaluate on all training data

=== Attribute Selection on all input data ===

Search Method:
    Attribute ranking.

Attribute Evaluator (unsupervised):
    Principal Components Attribute Transformer

Correlation matrix
  1     -0.58   0.65  -0.08  -0.37   0     -0.05
 -0.58   1     -0.44   0.06   0.23   0.05   0.07
  0.65  -0.44   1      0.38  -0.76   0.15  -0.15
 -0.08   0.06   0.38   1     -0.46   0.32  -0.07
 -0.37   0.23  -0.76  -0.46   1     -0.38   0.25
  0      0.05   0.15   0.32  -0.38   1     -0.11
 -0.05   0.07  -0.15  -0.07   0.25  -0.11   1   

eigenvalue    proportion    cumulative
  2.75383      0.3934       0.3934     -0.552TempMax+0.513HumedadRelativa-0.418HorasBrillo+0.335Nubosidad-0.27TempMin...
  1.58542      0.22649      0.61989    -0.521TempMin-0.484VeloViento-0.473Nubosidad+0.445HorasBrillo+0.225HumedadRelativa...
  0.9575       0.13679      0.75668    0.957Precipitacion+0.256TempMin+0.115TempMax+0.063HorasBrillo+0.038VeloViento...
  0.72674      0.10382      0.8605     -0.828VeloViento+0.447TempMin+0.21 TempMax+0.165Nubosidad-0.16HorasBrillo...
  0.51814      0.07402      0.93452    0.751Nubosidad-0.453TempMin-0.288HumedadRelativa+0.286HorasBrillo+0.239TempMax...
  0.31912      0.04559      0.98011    -0.608HorasBrillo-0.604HumedadRelativa-0.398TempMin-0.268Nubosidad+0.152Precipitacion...

Eigenvectors
 V1     V2     V3     V4     V5     V6   
-0.4184     0.4455     0.0629    -0.16       0.2865    -0.6084    HorasBrillo
 0.3352    -0.4734     0.0164     0.1648     0.7512    -0.2679    Nubosidad
-0.552      0.0616     0.1149     0.2104     0.2387     0.019     TempMax
-0.2699    -0.5214     0.2564     0.4468    -0.4526    -0.3976    TempMin
 0.5126     0.2248    -0.0011    -0.0795    -0.288     -0.6041    HumedadRelativa
-0.2088    -0.4839     0.0381    -0.8283    -0.0697    -0.1067    VeloViento
 0.1693     0.1307     0.9568    -0.1044     0.0634     0.1524    Precipitacion

Ranked attributes:
 0.6066  1 -0.552TempMax+0.513HumedadRelativa-0.418HorasBrillo+0.335Nubosidad-0.27TempMin...
 0.3801  2 -0.521TempMin-0.484VeloViento-0.473Nubosidad+0.445HorasBrillo+0.225HumedadRelativa...
 0.2433  3 0.957Precipitacion+0.256TempMin+0.115TempMax+0.063HorasBrillo+0.038VeloViento...
 0.1395  4 -0.828VeloViento+0.447TempMin+0.21 TempMax+0.165Nubosidad-0.16HorasBrillo...
 0.0655  5 0.751Nubosidad-0.453TempMin-0.288HumedadRelativa+0.286HorasBrillo+0.239TempMax...
 0.0199  6 -0.608HorasBrillo-0.604HumedadRelativa-0.398TempMin-0.268Nubosidad+0.152Precipitacion...

Selected attributes: 1,2,3,4,5,6 : 6


I would like to know how to interpret these results. I am also interested in knowing if I obtained good results and what other methods or tools I can use and how to interpret these results.

Attached file .arff

Thank you all.

_______________________________________________
Wekalist mailing list
Send posts to: [hidden email]
To subscribe, unsubscribe, etc., visit https://list.waikato.ac.nz/mailman/listinfo/wekalist
List etiquette: http://www.cs.waikato.ac.nz/~ml/weka/mailinglist_etiquette.html

Radiacion.arff (83K) Download Attachment
Reply | Threaded
Open this post in threaded view
|

Re: Query about classification of data in WEKA

Eibe Frank-2
Administrator
PCA is unsupervised, but you have a supervised (regression) problem and seem to be interested in the predictive power of the variables. The output is not really useful for what you are interested in. Use the “Classify” tab (or similar) to build and evaluate predictive models.

Consider taking a look at our introductory online courses on machine learning using WEKA:

  https://www.cs.waikato.ac.nz/ml/weka/courses.html

(You could use PCA to transform or reduce the data inside the regression model. There is a WEKA “filter” for that.)

Cheers,
Eibe

PS: Mailing list post are not sent back to the sender by default when people use the wekalist. There is a setting that you can fiddle with at the Waikato web page for the list to change that. Or check the archive for whether your message has been posted.

> On 6/02/2019, at 4:28 PM, Juan Sebastian Mejia <[hidden email]> wrote:
>
> Hello everyone, I'm new using WEKA. I am developing a solar radiation prediction project with some meteorological variables as input of the model: Hours of solar brightness, cloudiness, maximum temperature, minimum temperature, relative humidity, etc.
>
> I used WEKA as a tool to identify which of these input variables are the most relevant and correlated in relation to solar radiation.
>
> I chose Select Attributes: PrincipalComponents as an attribute evaluator and the Ranker search method.
>
> I got the following results:
>
> Evaluator:    weka.attributeSelection.PrincipalComponents -R 0.95 -A 5
> Search:       weka.attributeSelection.Ranker -T -1.7976931348623157E308 -N -1
> Relation:     solarradiation
> Instances:    1275
> Attributes:   8
>               HorasBrillo
>               Nubosidad
>               TempMax
>               TempMin
>               HumedadRelativa
>               VeloViento
>               Precipitacion
>               Radiacion
> Evaluation mode:    evaluate on all training data
>
> === Attribute Selection on all input data ===
>
> Search Method:
>     Attribute ranking.
>
> Attribute Evaluator (unsupervised):
>     Principal Components Attribute Transformer
>
> Correlation matrix
>   1     -0.58   0.65  -0.08  -0.37   0     -0.05
>  -0.58   1     -0.44   0.06   0.23   0.05   0.07
>   0.65  -0.44   1      0.38  -0.76   0.15  -0.15
>  -0.08   0.06   0.38   1     -0.46   0.32  -0.07
>  -0.37   0.23  -0.76  -0.46   1     -0.38   0.25
>   0      0.05   0.15   0.32  -0.38   1     -0.11
>  -0.05   0.07  -0.15  -0.07   0.25  -0.11   1    
>
> eigenvalue    proportion    cumulative
>   2.75383      0.3934       0.3934     -0.552TempMax+0.513HumedadRelativa-0.418HorasBrillo+0.335Nubosidad-0.27TempMin...
>   1.58542      0.22649      0.61989    -0.521TempMin-0.484VeloViento-0.473Nubosidad+0.445HorasBrillo+0.225HumedadRelativa...
>   0.9575       0.13679      0.75668    0.957Precipitacion+0.256TempMin+0.115TempMax+0.063HorasBrillo+0.038VeloViento...
>   0.72674      0.10382      0.8605     -0.828VeloViento+0.447TempMin+0.21 TempMax+0.165Nubosidad-0.16HorasBrillo...
>   0.51814      0.07402      0.93452    0.751Nubosidad-0.453TempMin-0.288HumedadRelativa+0.286HorasBrillo+0.239TempMax...
>   0.31912      0.04559      0.98011    -0.608HorasBrillo-0.604HumedadRelativa-0.398TempMin-0.268Nubosidad+0.152Precipitacion...
>
> Eigenvectors
>  V1     V2     V3     V4     V5     V6    
> -0.4184     0.4455     0.0629    -0.16       0.2865    -0.6084    HorasBrillo
>  0.3352    -0.4734     0.0164     0.1648     0.7512    -0.2679    Nubosidad
> -0.552      0.0616     0.1149     0.2104     0.2387     0.019     TempMax
> -0.2699    -0.5214     0.2564     0.4468    -0.4526    -0.3976    TempMin
>  0.5126     0.2248    -0.0011    -0.0795    -0.288     -0.6041    HumedadRelativa
> -0.2088    -0.4839     0.0381    -0.8283    -0.0697    -0.1067    VeloViento
>  0.1693     0.1307     0.9568    -0.1044     0.0634     0.1524    Precipitacion
>
> Ranked attributes:
>  0.6066  1 -0.552TempMax+0.513HumedadRelativa-0.418HorasBrillo+0.335Nubosidad-0.27TempMin...
>  0.3801  2 -0.521TempMin-0.484VeloViento-0.473Nubosidad+0.445HorasBrillo+0.225HumedadRelativa...
>  0.2433  3 0.957Precipitacion+0.256TempMin+0.115TempMax+0.063HorasBrillo+0.038VeloViento...
>  0.1395  4 -0.828VeloViento+0.447TempMin+0.21 TempMax+0.165Nubosidad-0.16HorasBrillo...
>  0.0655  5 0.751Nubosidad-0.453TempMin-0.288HumedadRelativa+0.286HorasBrillo+0.239TempMax...
>  0.0199  6 -0.608HorasBrillo-0.604HumedadRelativa-0.398TempMin-0.268Nubosidad+0.152Precipitacion...
>
> Selected attributes: 1,2,3,4,5,6 : 6
>
> I would like to know how to interpret these results. I am also interested in knowing if I obtained good results and what other methods or tools I can use and how to interpret these results.
>
> Attached file .arff
>
> Thank you all.
> <Radiacion.arff>_______________________________________________
> Wekalist mailing list
> Send posts to: [hidden email]
> To subscribe, unsubscribe, etc., visit 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]
To subscribe, unsubscribe, etc., visit 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
|

Re: Query about classification of data in WEKA

Juan Sebastian Mejia
Hello, thank you very much for replying.

Based on that answer I have other questions to ask:

When opening the .arff file in WEKA, apply a supervised filter: AttributeSelection. Then I used the “Classify” tab, specifically in the tree methods: RandomForest using training set and I obtained the following results:

"Scheme:       weka.classifiers.trees.RandomForest -P 100 -I 100 -num-slots 1 -K 0 -M 1.0 -V 0.001 -S 1
Relation:     solarradiation-weka.filters.supervised.attribute.AttributeSelection-Eweka.attributeSelection.CfsSubsetEval -P 1 -E 1-Sweka.attributeSelection.BestFirst -D 1 -N 5
Instances:    1275
Attributes:   3
              HorasBrillo
              VeloViento
              Radiacion
Test mode:    evaluate on training data

=== Classifier model (full training set) ===

RandomForest

Bagging with 100 iterations and base learner

weka.classifiers.trees.RandomTree -K 0 -M 1.0 -V 0.001 -S 1 -do-not-check-capabilities

Time taken to build model: 1.56 seconds

=== Evaluation on training set ===

Time taken to test model on training data: 0.35 seconds

=== Summary ===

Correlation coefficient                  0.9324
Mean absolute error                    362.3223
Root mean squared error                469.2963
Relative absolute error                 37.9267 %
Root relative squared error             39.2306 %
Total Number of Instances             1275
"    

My question is how can I interpret these results, and if these results are good that are useful for the project.

Thanks again for replying to my initial email.

El mar., 5 de feb. de 2019 a la(s) 22:39, Eibe Frank ([hidden email]) escribió:
PCA is unsupervised, but you have a supervised (regression) problem and seem to be interested in the predictive power of the variables. The output is not really useful for what you are interested in. Use the “Classify” tab (or similar) to build and evaluate predictive models.

Consider taking a look at our introductory online courses on machine learning using WEKA:

  https://www.cs.waikato.ac.nz/ml/weka/courses.html

(You could use PCA to transform or reduce the data inside the regression model. There is a WEKA “filter” for that.)

Cheers,
Eibe

PS: Mailing list post are not sent back to the sender by default when people use the wekalist. There is a setting that you can fiddle with at the Waikato web page for the list to change that. Or check the archive for whether your message has been posted.

> On 6/02/2019, at 4:28 PM, Juan Sebastian Mejia <[hidden email]> wrote:
>
> Hello everyone, I'm new using WEKA. I am developing a solar radiation prediction project with some meteorological variables as input of the model: Hours of solar brightness, cloudiness, maximum temperature, minimum temperature, relative humidity, etc.
>
> I used WEKA as a tool to identify which of these input variables are the most relevant and correlated in relation to solar radiation.
>
> I chose Select Attributes: PrincipalComponents as an attribute evaluator and the Ranker search method.
>
> I got the following results:
>
> Evaluator:    weka.attributeSelection.PrincipalComponents -R 0.95 -A 5
> Search:       weka.attributeSelection.Ranker -T -1.7976931348623157E308 -N -1
> Relation:     solarradiation
> Instances:    1275
> Attributes:   8
>               HorasBrillo
>               Nubosidad
>               TempMax
>               TempMin
>               HumedadRelativa
>               VeloViento
>               Precipitacion
>               Radiacion
> Evaluation mode:    evaluate on all training data
>
> === Attribute Selection on all input data ===
>
> Search Method:
>     Attribute ranking.
>
> Attribute Evaluator (unsupervised):
>     Principal Components Attribute Transformer
>
> Correlation matrix
>   1     -0.58   0.65  -0.08  -0.37   0     -0.05
>  -0.58   1     -0.44   0.06   0.23   0.05   0.07
>   0.65  -0.44   1      0.38  -0.76   0.15  -0.15
>  -0.08   0.06   0.38   1     -0.46   0.32  -0.07
>  -0.37   0.23  -0.76  -0.46   1     -0.38   0.25
>   0      0.05   0.15   0.32  -0.38   1     -0.11
>  -0.05   0.07  -0.15  -0.07   0.25  -0.11   1   
>
> eigenvalue    proportion    cumulative
>   2.75383      0.3934       0.3934     -0.552TempMax+0.513HumedadRelativa-0.418HorasBrillo+0.335Nubosidad-0.27TempMin...
>   1.58542      0.22649      0.61989    -0.521TempMin-0.484VeloViento-0.473Nubosidad+0.445HorasBrillo+0.225HumedadRelativa...
>   0.9575       0.13679      0.75668    0.957Precipitacion+0.256TempMin+0.115TempMax+0.063HorasBrillo+0.038VeloViento...
>   0.72674      0.10382      0.8605     -0.828VeloViento+0.447TempMin+0.21 TempMax+0.165Nubosidad-0.16HorasBrillo...
>   0.51814      0.07402      0.93452    0.751Nubosidad-0.453TempMin-0.288HumedadRelativa+0.286HorasBrillo+0.239TempMax...
>   0.31912      0.04559      0.98011    -0.608HorasBrillo-0.604HumedadRelativa-0.398TempMin-0.268Nubosidad+0.152Precipitacion...
>
> Eigenvectors
>  V1     V2     V3     V4     V5     V6   
> -0.4184     0.4455     0.0629    -0.16       0.2865    -0.6084    HorasBrillo
>  0.3352    -0.4734     0.0164     0.1648     0.7512    -0.2679    Nubosidad
> -0.552      0.0616     0.1149     0.2104     0.2387     0.019     TempMax
> -0.2699    -0.5214     0.2564     0.4468    -0.4526    -0.3976    TempMin
>  0.5126     0.2248    -0.0011    -0.0795    -0.288     -0.6041    HumedadRelativa
> -0.2088    -0.4839     0.0381    -0.8283    -0.0697    -0.1067    VeloViento
>  0.1693     0.1307     0.9568    -0.1044     0.0634     0.1524    Precipitacion
>
> Ranked attributes:
>  0.6066  1 -0.552TempMax+0.513HumedadRelativa-0.418HorasBrillo+0.335Nubosidad-0.27TempMin...
>  0.3801  2 -0.521TempMin-0.484VeloViento-0.473Nubosidad+0.445HorasBrillo+0.225HumedadRelativa...
>  0.2433  3 0.957Precipitacion+0.256TempMin+0.115TempMax+0.063HorasBrillo+0.038VeloViento...
>  0.1395  4 -0.828VeloViento+0.447TempMin+0.21 TempMax+0.165Nubosidad-0.16HorasBrillo...
>  0.0655  5 0.751Nubosidad-0.453TempMin-0.288HumedadRelativa+0.286HorasBrillo+0.239TempMax...
>  0.0199  6 -0.608HorasBrillo-0.604HumedadRelativa-0.398TempMin-0.268Nubosidad+0.152Precipitacion...
>
> Selected attributes: 1,2,3,4,5,6 : 6
>
> I would like to know how to interpret these results. I am also interested in knowing if I obtained good results and what other methods or tools I can use and how to interpret these results.
>
> Attached file .arff
>
> Thank you all.
> <Radiacion.arff>_______________________________________________
> Wekalist mailing list
> Send posts to: [hidden email]
> To subscribe, unsubscribe, etc., visit 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]
To subscribe, unsubscribe, etc., visit 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]
To subscribe, unsubscribe, etc., visit 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
|

Re: Query about classification of data in WEKA

Eibe Frank-2
Administrator

The performance statistics indicate that the random forest model can quite accurately predict the target attribute’s values given the values of the other two attributes, e.g., the Pearson correlation between actual and predicted values is close to one.

 

However, you have applied supervised attribute selection (CFS) outside the cross-validation loop, so your results will be optimistically biased. You need to use the AttributeSelectedClassifier in the Classify panel to avoid this.

 

Please take a look at the free introductory courses on data mining with WEKA at

 

http://www.cs.waikato.ac.nz/ml/weka

 

for more information on the key concepts that you need to know about.

 

Cheers,

Eibe

 

From: [hidden email]
Sent: Thursday, 7 February 2019 4:04 PM
To: [hidden email]
Subject: Re: [Wekalist] Query about classification of data in WEKA

 

Hello, thank you very much for replying.

Based on that answer I have other questions to ask:

When opening the .arff file in WEKA, apply a supervised filter: AttributeSelection. Then I used the “Classify” tab, specifically in the tree methods: RandomForest using training set and I obtained the following results:

 

"Scheme:       weka.classifiers.trees.RandomForest -P 100 -I 100 -num-slots 1 -K 0 -M 1.0 -V 0.001 -S 1
Relation:     solarradiation-weka.filters.supervised.attribute.AttributeSelection-Eweka.attributeSelection.CfsSubsetEval -P 1 -E 1-Sweka.attributeSelection.BestFirst -D 1 -N 5
Instances:    1275
Attributes:   3
              HorasBrillo
              VeloViento
              Radiacion
Test mode:    evaluate on training data

=== Classifier model (full training set) ===

RandomForest

Bagging with 100 iterations and base learner

weka.classifiers.trees.RandomTree -K 0 -M 1.0 -V 0.001 -S 1 -do-not-check-capabilities

Time taken to build model: 1.56 seconds

=== Evaluation on training set ===

Time taken to test model on training data: 0.35 seconds

=== Summary ===

Correlation coefficient                  0.9324
Mean absolute error                    362.3223
Root mean squared error                469.2963
Relative absolute error                 37.9267 %
Root relative squared error             39.2306 %
Total Number of Instances             1275
"    

My question is how can I interpret these results, and if these results are good that are useful for the project.

 

Thanks again for replying to my initial email.

 

El mar., 5 de feb. de 2019 a la(s) 22:39, Eibe Frank ([hidden email]) escribió:

PCA is unsupervised, but you have a supervised (regression) problem and seem to be interested in the predictive power of the variables. The output is not really useful for what you are interested in. Use the “Classify” tab (or similar) to build and evaluate predictive models.

Consider taking a look at our introductory online courses on machine learning using WEKA:

  https://www.cs.waikato.ac.nz/ml/weka/courses.html

(You could use PCA to transform or reduce the data inside the regression model. There is a WEKA “filter” for that.)

Cheers,
Eibe

PS: Mailing list post are not sent back to the sender by default when people use the wekalist. There is a setting that you can fiddle with at the Waikato web page for the list to change that. Or check the archive for whether your message has been posted.


> On 6/02/2019, at 4:28 PM, Juan Sebastian Mejia <[hidden email]> wrote:
>
> Hello everyone, I'm new using WEKA. I am developing a solar radiation prediction project with some meteorological variables as input of the model: Hours of solar brightness, cloudiness, maximum temperature, minimum temperature, relative humidity, etc.
>
> I used WEKA as a tool to identify which of these input variables are the most relevant and correlated in relation to solar radiation.
>
> I chose Select Attributes: PrincipalComponents as an attribute evaluator and the Ranker search method.
>
> I got the following results:
>
> Evaluator:    weka.attributeSelection.PrincipalComponents -R 0.95 -A 5
> Search:       weka.attributeSelection.Ranker -T -1.7976931348623157E308 -N -1
> Relation:     solarradiation
> Instances:    1275
> Attributes:   8
>               HorasBrillo
>               Nubosidad
>               TempMax
>               TempMin
>               HumedadRelativa
>               VeloViento
>               Precipitacion
>               Radiacion
> Evaluation mode:    evaluate on all training data
>
> === Attribute Selection on all input data ===
>
> Search Method:
>     Attribute ranking.
>
> Attribute Evaluator (unsupervised):
>     Principal Components Attribute Transformer
>
> Correlation matrix
>   1     -0.58   0.65  -0.08  -0.37   0     -0.05
>  -0.58   1     -0.44   0.06   0.23   0.05   0.07
>   0.65  -0.44   1      0.38  -0.76   0.15  -0.15
>  -0.08   0.06   0.38   1     -0.46   0.32  -0.07
>  -0.37   0.23  -0.76  -0.46   1     -0.38   0.25
>   0      0.05   0.15   0.32  -0.38   1     -0.11
>  -0.05   0.07  -0.15  -0.07   0.25  -0.11   1   
>
> eigenvalue    proportion    cumulative
>   2.75383      0.3934       0.3934     -0.552TempMax+0.513HumedadRelativa-0.418HorasBrillo+0.335Nubosidad-0.27TempMin...
>   1.58542      0.22649      0.61989    -0.521TempMin-0.484VeloViento-0.473Nubosidad+0.445HorasBrillo+0.225HumedadRelativa...
>   0.9575       0.13679      0.75668    0.957Precipitacion+0.256TempMin+0.115TempMax+0.063HorasBrillo+0.038VeloViento...
>   0.72674      0.10382      0.8605     -0.828VeloViento+0.447TempMin+0.21 TempMax+0.165Nubosidad-0.16HorasBrillo...
>   0.51814      0.07402      0.93452    0.751Nubosidad-0.453TempMin-0.288HumedadRelativa+0.286HorasBrillo+0.239TempMax...
>   0.31912      0.04559      0.98011    -0.608HorasBrillo-0.604HumedadRelativa-0.398TempMin-0.268Nubosidad+0.152Precipitacion...
>
> Eigenvectors
>  V1     V2     V3     V4     V5     V6   
> -0.4184     0.4455     0.0629    -0.16       0.2865    -0.6084    HorasBrillo
>  0.3352    -0.4734     0.0164     0.1648     0.7512    -0.2679    Nubosidad
> -0.552      0.0616     0.1149     0.2104     0.2387     0.019     TempMax
> -0.2699    -0.5214     0.2564     0.4468    -0.4526    -0.3976    TempMin
>  0.5126     0.2248    -0.0011    -0.0795    -0.288     -0.6041    HumedadRelativa
> -0.2088    -0.4839     0.0381    -0.8283    -0.0697    -0.1067    VeloViento
>  0.1693     0.1307     0.9568    -0.1044     0.0634     0.1524    Precipitacion
>
> Ranked attributes:
>  0.6066  1 -0.552TempMax+0.513HumedadRelativa-0.418HorasBrillo+0.335Nubosidad-0.27TempMin...
>  0.3801  2 -0.521TempMin-0.484VeloViento-0.473Nubosidad+0.445HorasBrillo+0.225HumedadRelativa...
>  0.2433  3 0.957Precipitacion+0.256TempMin+0.115TempMax+0.063HorasBrillo+0.038VeloViento...
>  0.1395  4 -0.828VeloViento+0.447TempMin+0.21 TempMax+0.165Nubosidad-0.16HorasBrillo...
>  0.0655  5 0.751Nubosidad-0.453TempMin-0.288HumedadRelativa+0.286HorasBrillo+0.239TempMax...
>  0.0199  6 -0.608HorasBrillo-0.604HumedadRelativa-0.398TempMin-0.268Nubosidad+0.152Precipitacion...
>
> Selected attributes: 1,2,3,4,5,6 : 6
>
> I would like to know how to interpret these results. I am also interested in knowing if I obtained good results and what other methods or tools I can use and how to interpret these results.
>
> Attached file .arff
>
> Thank you all.
> <Radiacion.arff>_______________________________________________
> Wekalist mailing list
> Send posts to: [hidden email]
> To subscribe, unsubscribe, etc., visit 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]
To subscribe, unsubscribe, etc., visit 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]
To subscribe, unsubscribe, etc., visit https://list.waikato.ac.nz/mailman/listinfo/wekalist
List etiquette: http://www.cs.waikato.ac.nz/~ml/weka/mailinglist_etiquette.html