SMO with kernel

classic Classic list List threaded Threaded
1 message Options
Reply | Threaded
Open this post in threaded view
|

SMO with kernel

alaa.3j
Hi

I am doing two experiments using SMO classifier with PolyKernel and
RBFKernel. My data is changed every hour.

My two questions are:

1- The result of PolyKernel is a linear equation with + and -. Does +,-
means that with data change the positive attribute, positively affect the
equation and negative attribute, decreases over time and negatively affect
the equation?
PolyKernel result:
weights (not support vectors):
 +       0.0011 * (normalized) Video age
 -       0.0049 * (normalized) Category ID
 +       0.3531 * (normalized) Channel view
 -       0.0612 * (normalized) Channel subscriber
 -       0.0898 * (normalized) \# Likes
 +       0.7052 * (normalized) \# Dislike
 +       0.1695 * (normalized) \# Comments Total
 -       0.0019 * (normalized) Comment sentiment polarity
 -       0.0033 * (normalized) Comment sentiment score
 -       0.0258

2- The result of RBFKernel is something looks like below, I want to get the
weights as linear equations is this possible?
Support vectors:
+1.0 * k[5]
+0.22039585186952054 * k[6]
+1.0 * k[8]
+0.11247675327160575 * k[18]
+0.22166880203736 * k[19]
+0.4693679132769697 * k[21]
-0.09589108727704014 * k[23]
+0.005582406477885327 * k[24]
-1.0 * k[37]
+1.0 * k[38]
+1.0 * k[39]
-1.0 * k[40]
+1.0 * k[41]
+1.0 * k[42]
-0.048643049579471784 * k[43]
 +       0.117

Thanks



--
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