Fig 1.
Generation of documents through topics following the Dirichlet distribution.
Fig 2.
LDA training and inference phases. D represents the total number of documents of the corpus and N the number of topics.
Fig 3.
Workflow and data transfer scheme.
Table 1.
LDA attributes.
Fig 4.
Select the filter from the filter drop-down list.
Fig 5.
Modify the default value of the filter parameters.
Fig 6.
Result of the filter application.
An attribute is generated for each topic.
Table 2.
Classifier result.
Table 3.
Best parameter values.
Table 4.
Experiments results.
Fig 7.
Kappa results with Naive Bayes.
Ten fold cross validation mean Kappa results for Naive Bayes Classifier.
Fig 8.
CPU usage time in milliseconds for SVM classifier.
Fig 9.
CPU usage time in milliseconds for k-NN classifier.
Fig 10.
CPU usage time with Naive Bayes.
CPU usage time in milliseconds for Naive Bayes classifier.