Sentiment Analyzer with RapidMiner that
- Collects latest film reviews from a website (scraping)
- Process the reviews. Tokenizing, removing stopwords and outliers and stemming.
- Analyze data and generated a Naive Bayes and a Support Vector Machine model.
- Using the model generated in the last step, predicts the sentiment of new films reviews.
Holds part of data that will be processed
The pang_lee folder contains the training and test data
The filmaffinity folder contains manually downloaded data to do predictions
Contains all the RapidMiner files.
The generator process takes the data from the data folder and trains and tests a model
The optimizer process iterates the same model with differents parameters to find the best one
The scraper process downloads new film reviews from www.rottentomatoes.com to do predictions
The predictor_rottentomatoes does predictions to the downloaded film reviews from that page
The predictor_filmaffinity does predictions to the manually downloaded film reviews from the filmaffinity data folder