luhh17 / sentiment_dataset

Credits

This repository was taken from https://github.com/CS287/HW1/tree/master/data
We modify the preprocessing script to get pandas output format.

Usage

./preprocess.py [-m yaml_file] CORPUS
# Use -m option to specify the loading setting yaml file.
# Example: ./preprocess.py -m data.yaml MR
# A file named MR.pkl will be generated (pandas format)
# Please see comments in corpus.yaml to see how to configure the setting

Load data with python

corpus = pandas.read_pickle('MR.pkl')
sentences, labels = list(corpus.sentence), list(corpus.label)

Datasets

Data	Classes	Average sentence length	Dataset size	Vocab size	Number of words present in word2vec	Test size
MR	2	20	10662	18765	16448	CV
SST1	5	18	11855	17836	16262	2210
SST2	2	19	9613	16185	14838	1821
Subj	2	23	10000	21323	17913	CV
TREC	6	10	5952	9592	9125	500
CR	2	19	3775	5340	5046	CV
MPQA	2	3	10606	6246	6083	CV

The following datasets are included in this directory:

• MR: Movie reviews with one sentence per review. Classification involves detecting positive/negative reviews (Pang and Lee, 2005). Link

• SST-1: Stanford Sentiment Treebank - an extension of MR but with train/dev/test splits provided and fine-grained labels (very positive, positive, neutral, negative, very negative), re-labeled by Socher et al. (2013). Link

  Note that data is actually provided at the phrase-level and hence we train the model on both phrases and sentences but only score on sentences at test time, as in Socher et al. (2013), Kalchbrenner et al. (2014), and Le and Mikolov (2014). Thus the training set is an order of magnitude larger than listed in the above table.

• SST-2 Same as SST-1 but with neutral reviews removed and binary labels.

• Subj: Subjectivity dataset where the task is to classify a sentence as being subjective or objective (Pang and Lee, 2004). Link

• TREC: TREC question dataset - task involves classifying a question into 6 question types (whether the question is about person, location, numeric information, etc.) (Li and Roth, 2002). Link

• CR: Customer reviews of various products (cameras, MP3s etc.). Task is to predict positive/negative reviews (Hu and Liu, 2004). Link