Filter sports for days. Labeled data continuously scraped from news and Twitter sources. Optimized for fewer false positives (classifying non-sports as sports) at the expense of missing more sports than necessary, as deleting non-sports content is more detrimental than not deleting sports content. Current stats:
> python classify.py
Training set size: 192044
MultinomialNB
precision recall f1-score support
-1 0.96 0.98 0.97 33749
1 0.97 0.96 0.97 30265
avg / total 0.97 0.97 0.97 64014
[[32944 805]
[ 1200 29065]]
LogisticRegression
precision recall f1-score support
-1 0.97 0.97 0.97 33749
1 0.97 0.97 0.97 30265
avg / total 0.97 0.97 0.97 64014
[[32717 1032]
[ 932 29333]]
Ensemble at .9 threshold:
precision recall f1-score support
-1 0.92 0.99 0.96 33749
1 0.99 0.90 0.95 30265
avg / total 0.95 0.95 0.95 64014
[[33510 239]
[ 2882 27383]]
