sentiment: Conduct dictionary-based sentiment analysis.

In markallenthornton/affectr: Weighted Dictionary Analysis of Text Affect

Description

Tokenizes character vectors, counts words, and then takes weighted average to estimate sentiment.

Usage

sentiment(x,tokens,scores)

Arguments

x	length n character vector. If you're input dictionary is lower case only, then be sure to preprocess inputs to x using tolower.
tokens	length m character vector containing dictionary words
scores	m x d numerical matrix containing scores for each dictionary word on d-dimensions. Values can be weighted continuous scores or binary.

Value

n x d numerical matrix with scores on x for dimensions d

See Also

count_tokens tokenize

Examples

dictionary <- c("great","good","bad","awful")
scores <- matrix(c(1,.5,-.5,-1),4,1)
sentiment("This package is great.",dictionary,scores)
sentiment("This package is good.",dictionary,scores)
sentiment("This package is bad.",dictionary,scores)
sentiment("This package is awful.",dictionary,scores)

markallenthornton/affectr documentation built on May 17, 2019, 2:15 a.m.