“如何减少模式l size - delete where weight
Contributor II
Hi there.. me again
I have a process to create a textmining model. My model is too big so I want it to use data where weight>0... on the weight table I see lots of words with weight=0 that I want to delete - not include in the model. Is there a way to do this?
thanks again for your help!
在这里我的代码
I have a process to create a textmining model. My model is too big so I want it to use data where weight>0... on the weight table I see lots of words with weight=0 that I want to delete - not include in the model. Is there a way to do this?
thanks again for your help!
在这里我的代码
Tagged:
0
Answers
you could use a weighting scheme before applying the learner, this would reduce the number of attributes and hence the length of support vectors. A similar weighting to the svm's weight vectors will be given by the SVMWeighting operator. If you need to apply the weights lateron, you could use the attributeWeightsApplier.
Greetings,
Sebastian