createAmazonSentiment: Create Amazon Sentiment dataset

In jkrijthe/createdatasets: Download and preprocess common benchmark dataset for use in reproduceable simulation studies

Description

The goal in this dataset is to predict binary (positive/negative) sentiment from product reviews taken from Amazon.com in four different categories: books, dvds, electronics and kitchen.

Usage

createAmazonSentiment(file = getfilepath("amazonsentiment.rds"),
  write = TRUE, read = TRUE)

Arguments

file	character; path/filename to write data file to
write	logical; should the dataset be written to disk for later use? (default: TRUE)
read	logical; should we try to read the dataset from the specified location first? (default: TRUE)

Details

For 27677 reviews, 1110349 unigram and bigram features are given in a sparse matrix X, along with a label vector y and a vector indicating the domain. Note the data is NOT a data.table, but sparse matrix generated by sparseMatrix.

Task: Classification: Use X to predict y, possibly in a domain adaptation setting.

Value

List containing:

• "X" dgCMatrix; sparse matrix with count of unigram and bigram features

• "y" factor; labels

• "domains" factor; domain/category for each review

See Also

createAmazonSentimentStars, http://www.cs.jhu.edu/~mdredze/datasets/sentiment/

jkrijthe/createdatasets documentation built on May 19, 2019, 12:44 p.m.