#!/usr/bin/env python """Compute VSPS sentiment scores; store in the database.""" import extract import score import publish scorer = score.SentimentScorer.from_vaccine_phrases() results = [(_id, scorer.get_document_score(text, normalize=False)) for (_id, text) in extract.extract_text()] publish.publish_sentiment('vsps', results) print 'published %d results' % len(results)
#!/usr/bin/env python # Andrew Whitaker """Compute sentiment scores using logistic regression; store in the database.""" import sys # Hack: append common/ to sys.path sys.path.append("../common") import extract import publish from scikit_scorer import * from insert_dict import * import numpy as np scorer = ScikitScorer(create_logistic_regression_classifier()) results = [] for i, (tweet_id, text) in enumerate(extract.extract_text('tweets_2014')): score = np.asscalar(scorer.get_document_score(text)) results.append((tweet_id, score)) if (i % 1000) == 0: print "%d" % i publish.publish_sentiment('logistic', results) print 'published %d results' % len(results)
#!/usr/bin/env python """Compute naive bayes sentiment scores; store in the database.""" import extract import publish from scikit_scorer import * import numpy as np scorer = ScikitScorer(create_naive_bayes_classifier()) results = [] for i, (_id, text) in enumerate(extract.extract_text()): score = np.asscalar(scorer.get_document_score(text)) results.append((_id, score)) if (i % 1000) == 0: print "%d" % i publish.publish_sentiment('naivebayes', results) print 'published %d results' % len(results)
#!/usr/bin/env python """Compute afinn111 sentiment scores; store in the database.""" import extract import score import publish scorer = score.SentimentScorer.from_afinn_111() results = [(_id, scorer.get_document_score(text, normalize=False)) for (_id, text) in extract.extract_text()] # Revised == without the "flu" keyword publish.publish_sentiment('afinn111_revised', results) print 'published %d results' % len(results)