diff --git a/README.md b/README.md index 52c9de1..b290c26 100644 --- a/README.md +++ b/README.md @@ -1,5 +1,5 @@ # Tweet-Semantics -Three distinct sentiment analysis classifier are used that are capable of labellign tweets as either positive, neutral or negative. The tweet dataset and general project was heavily inspired by the semeval competition. +Three distinct sentiment analysis classifier are used that are capable of labelling tweets as either positive, neutral or negative. The tweet dataset and general project was heavily inspired by the semeval competition. The classifiers were tested against evaluated according to the macro-averaged F1-score, which meant that the inbalance in the tweets labels (negatives were nearly half of positives) was quite an important obstacle.