This system is an implementation of a Convolutional Neural Network to extract sentiments of texts in Brazilian Portuguese language.
The model and architecture is based on Kim (2014) work, several modifications had to be made in order to adapt for Brazilian Portuguese texts and to get the output and performance desired.
- TensorFlow is the deep learning framework used to train, test and validate the model.
- HPCC Platform is used to pre-process text data into the format that the model expects.
This work was used as my Final Paper project in order to receive the Computer Engineering academic degree at State University of Ponta Grossa (UEPG).
The academic paper from this work is available here (document in Portuguese): Análise de Sentimentos em Textos em Português do Brasil Usando Redes Neurais Convolucionais.
Feel free to contact me at any time if you have questions or wanna talk about this work. You are free to use and reference the model here presented as long as you give credits to the author.
- Author: Lucas Alves Sobrinho
- Email: lasobrinho7@gmail.com