Source code for the paper publishe in Revista Democracia Digital e Governo Eletrônico.
Link for the paper: https://buscalegis.ufsc.br/revistas/index.php/observatoriodoegov/article/view/316
In this subproject, we add the pipeline used for text classification using Orange 3. The file can be opened as a project in the Orange 3 tool.
We used the following classical ML techniques:
- k-Nearest Neighbors
- Random Forest
- Naïve Bayes
- feed-forward Neural Networks
- Logistic Regression
- Support Vector Machine
- Neural Network
In this subproject, we present the code of the pipeline for text classification implemented in the Python language and open-source tools, such as, NLTK, Scikit-Learn, and Pandas.
The following DL techniques were used:
- LSTM
- CNN (based on Kim (2014))
- Bi-LSTM with Self-Attention Chalkidis, Kampas (2019)
- We did not make our datasets for text classification available due to the existing personal data in the documents.