Brazilian parties reflect senators opinions?
Clusters, Senators, Parties, Speech, Natural Language Processing
The research project aims to verify wheter senators from the same party have similar speeches. Using natural language processing techniques, the speeches will be cleaned with pre-processing techniques, later they will be separated into clusters. Senators will be grouped according to the clusters of your speeches. The clusters will be compared with the division of senators into parties. This will verify whether senators from the same party tend to be in the same cluster. It will also be verified which words represent the most of the clusters’ speeches. This helps in the perception of whether the clusters are related to party positioning (left, center and right). The model used separates the speeches by legislature, speeches from the 51th to the 55th legislature will be analyzed.