PDA

View Full Version : Weka In Web Application How?



JauB
02-07-2008, 03:40 PM
Hi,
is there a way to use WEKA by using the browser or we must install it in a local post to use it?

Mark
02-07-2008, 04:29 PM
Pentaho Data Mining (Weka) at present operates as an application. There are some projects that have looked at exposing some of Weka's functionality as a web service. Take a look at:

http://grid.deis.unical.it/weka4ws/
http://users.cs.cf.ac.uk/Ali.Shaikhali/faehim/
https://list.scms.waikato.ac.nz/mailman/htdig/wekalist/2008-January/012231.html

Cheers,
Mark.

JauB
02-08-2008, 05:19 AM
So if i understand i can't interface WEKA with un other web application because WEKA must be installed in the local post.
the probleme is that iwe have a web application which we want to interface it with PENTAHO (for DATA MINING), our application is used by 400 persons and i we can't install weka in 400 PC, this is my reason to think in WEKA in the web.
do you have an idea for my probleme.
thks

Mark
02-08-2008, 03:49 PM
This is correct. Integration of Weka into the Pentaho platform will continue this year. At present, there is some integration with Pentaho Data Integration (Kettle) - there is the ability to use pre-trained Weka classifiers or clusters in a Kettle transform to score data. This functionality is available through the Pentaho platform.

At some later stage there will be the ability to access Weka functionality via a web interface.

There is one other project that allows simple data mining reports to be generated using Weka via an online server. See:

http:// www.auknomi.com

Cheers,
Mark.

JauB
02-11-2008, 12:19 PM
This is correct. Integration of Weka into the Pentaho platform will continue this year. At present, there is some integration with Pentaho Data Integration (Kettle) - there is the ability to use pre-trained Weka classifiers or clusters in a Kettle transform to score data. This functionality is available through the Pentaho platform.

At some later stage there will be the ability to access Weka functionality via a web interface.

There is one other project that allows simple data mining reports to be generated using Weka via an online server. See:

http:// www.auknomi.com (http://www.auknomi.com)

Cheers,
Mark.

can we establish a tree decision with weka?

Mark
02-11-2008, 05:07 PM
Yes, Weka includes various algorithms for constructing decision trees, including an implementation of C4.5 (J48).

Cheers,
Mark.

JauB
02-12-2008, 06:13 AM
Yes, Weka includes various algorithms for constructing decision trees, including an implementation of C4.5 (J48).

Cheers,
Mark.
thanks,
when i use an exemple of weka's examples like contact-lenses.arff ( .arff file) it works when i try to visualize the tree. but when i use a CVS file for example my tree has no node but just one line like :
sunny;75;70;yes;play(14.0/13.0)
i don't understand what is it meaning and why i have no node!
am i obliged to work just with a arff file for visualizing the tree or what?

in the classifier output i got :

=== Run information ===
Scheme: weka.classifiers.trees.J48 -C 0.25 -M 2
Relation: weather2
Instances: 14
Attributes: 1
Outlook;Temp;Humidity;Windy;Class
Test mode: evaluate on training data
=== Classifier model (full training set) ===
J48 pruned tree
------------------
: sunny;75;70;yes;Play (14.0/13.0)
Number of Leaves : 1
Size of the tree : 1

Time taken to build model: 0 seconds
=== Evaluation on training set ===
=== Summary ===
Correctly Classified Instances 1 7.1429 %
Incorrectly Classified Instances 13 92.8571 %
Kappa statistic 0
Mean absolute error 0.1327
Root mean squared error 0.2575
Relative absolute error 100 %
Root relative squared error 100 %
Total Number of Instances 14
=== Detailed Accuracy By Class ===
TP Rate FP Rate Precision Recall F-Measure Class
1 1 0.071 1 0.133 sunny;75;70;yes;Play
0 0 0 0 0 sunny;80;90;yes;DontPlay
0 0 0 0 0 sunny;85;85;no;DontPlay
0 0 0 0 0 sunny;72;95;no;DontPlay
0 0 0 0 0 sunny;69;70;no;Play
0 0 0 0 0 overcast;72;90;yes;Play
0 0 0 0 0 overcast;83;78;no;Play
0 0 0 0 0 overcast;64;65;yes;Play
0 0 0 0 0 overcast;81;75;no;Play
0 0 0 0 0 rain;71;80;yes;DontPlay
0 0 0 0 0 rain;65;70;yes;DontPlay
0 0 0 0 0 rain;75;80;no;Play
0 0 0 0 0 rain;68;80;no;Play
0 0 0 0 0 rain;70;96;no;Play
=== Confusion Matrix ===
a b c d e f g h i j k l m n <-- classified as
1 0 0 0 0 0 0 0 0 0 0 0 0 0 | a = sunny;75;70;yes;Play
1 0 0 0 0 0 0 0 0 0 0 0 0 0 | b = sunny;80;90;yes;DontPlay
1 0 0 0 0 0 0 0 0 0 0 0 0 0 | c = sunny;85;85;no;DontPlay
1 0 0 0 0 0 0 0 0 0 0 0 0 0 | d = sunny;72;95;no;DontPlay
1 0 0 0 0 0 0 0 0 0 0 0 0 0 | e = sunny;69;70;no;Play
1 0 0 0 0 0 0 0 0 0 0 0 0 0 | f = overcast;72;90;yes;Play
1 0 0 0 0 0 0 0 0 0 0 0 0 0 | g = overcast;83;78;no;Play
1 0 0 0 0 0 0 0 0 0 0 0 0 0 | h = overcast;64;65;yes;Play
1 0 0 0 0 0 0 0 0 0 0 0 0 0 | i = overcast;81;75;no;Play
1 0 0 0 0 0 0 0 0 0 0 0 0 0 | j = rain;71;80;yes;DontPlay
1 0 0 0 0 0 0 0 0 0 0 0 0 0 | k = rain;65;70;yes;DontPlay
1 0 0 0 0 0 0 0 0 0 0 0 0 0 | l = rain;75;80;no;Play
1 0 0 0 0 0 0 0 0 0 0 0 0 0 | m = rain;68;80;no;Play
1 0 0 0 0 0 0 0 0 0 0 0 0 0 | n = rain;70;96;no;Play

thnks lot

Mark
02-13-2008, 07:59 PM
Try using commas to separate your columns, rather than semicolons.

Cheers,
Mark.

JauB
02-14-2008, 08:28 AM
Try using commas to separate your columns, rather than semicolons.

Cheers,
Mark.

thanks it works.