Text sentiment classification

2017 · Machine Learning

Github Repo

Problem description:

Use an RNN model to determine whether a Twitter post is positive or negative, label 1 means positive and label 0 means negative.

Word embedding

• Machine learns the meaning of words from reading a lot of documents without supervision
• word embedding method used in the project: word2vec model

Compare the scores between a bag of words (BOW) and a RNN model

	“today is a good day, but it is hot”	“today is hot, but it is a good day”
BOW	0.49039897	0.48992708
RNN	0.31660247	0.94093776

• BOW only cares about the presence of the given vocabularies but not the order
• The score of the RNN model depends more on the order

Semi-supervised learning

• An approach to machine learning that combines a small amount of labeled texts with a large amount of unlabeled texts during training

Result

• Achieved 262/344 (Top $77\%$) rank in the Kaggle competition