Help with Text Classification with BayesNet
Contributor IHello, I am trying to classify some questions into answerable or not.
It works with the techniques ( Decision Tree, Naive Bayes, Random Forest).
But with W-BayesNet and Neural Net it is taking too long and java stops processing after some time.
When the sample is at 1000 questions, the process executes in about 1 minute. More than that, it never ends.
I need some help with it.
The Process:
1.Retrieve
2.Sample (I'd like 3000)
3.Select Attributtes
4.Nominal to Text
5. Process Documents From Data
5.1 Extract Content 5.2 Transform Cases 5.3 Tokenize 5.4 Filter Stopwords 5.5 Stem 5.6 Filter Tokens
6. Select Attributes (no missing values)
7. Set Role
8. Cross Validation
8.1 Training - W-BayesNet
8.2 Testing - Apply Model and Performance
The XML from W-BayesNet Process is attached.
