Connotation is a project mainly written in PYTHON and PHP, it's free.
Connotation analysis in short messages
Dependencies:
Current usage: python main.py
Configuration is done from the settings.cfg file.
The current configuration uses a dummy algorithm to calculate scores of reviews drawn from IMDB. The rating associated with the review is included in the output to be able to infer statistics from the output.
The interface to SentiWordNet first contacts WordNet to extract all the synsets associated with a word. Then, SentiWordNet is consulted to draw the scores for every synset.
The algorithm first makes an average of all scores associated with the synsets for a given word. Then, it makes an average out of all the scores for a given text.
The data for the IMDB reviews was scraped with a PHP script (see datasource/imdb-reviews). The script inserts the data into a mysql-database, from which the python code draws them from.
The analysis/imdb.out file contains the output of the run over all reviews. When running analysis/imdb.py over the output file, you will see the percentage of correct evaluations and the average score associated with each rating.