Dictionary-Based Sentiment Analysis

Performs a sentiment analysis of textual contents in R. This implementation utilizes various existing dictionaries, such as Harvard IV, or finance-specific dictionaries. Furthermore, it can also create customized dictionaries. The latter uses LASSO regularization as a statistical approach to select relevant terms based on an exogenous response variable.


Sentiment Analysis

Build Status CRAN_Status_Badge

SentimentAnalysis performs a sentiment analysis of textual contents in R. This implementation utilizes various existing dictionaries, such as QDAP, Harvard IV or Loughran-McDonald. Furthermore, it can also create customized dictionaries. The latter uses LASSO regularization as a statistical approach to select relevant terms based on an exogenous response variable.

Overview

The most important functions in SentimentAnalysis are:

  • Compute sentiment scores from contents stored in different formats with analyzeSentiment().

  • If desired, convert the continuous scores to either binary sentiment classes (negative or positive) or tertiary directions (negative, neutral or positive). This conversion can be done with convertToBinary() or convertToDirection() respectively.

  • Compare the calculated sentiment socres with a baseline (i.e. a gold standard). Here, compareToResponse() performs a statistical evaluation, while plotSentimentResponse() enables a visual comparison.

  • Generate customized dictionaries with the help of generateDictionary() as part of an advanced analysis. However, this prerequisites a response variable (i.e. the baseline).

To see examples of these functions in use, check out the help pages, the demos and the vignette.

Usage

This section shows the basic functionality of how to perform a sentiment analysis. First, install the package from CRAN. Then load the corresponding package SentimentAnalysis.

 
library(SentimentAnalysis)

Quick demonstration

This simple example shows how to perform a sentiment analysis of a single string. The result is a two-level factor with levels "positive" and "negative."

 
# Analyze a single string to obtain a binary response (positive / negative)
sentiment <- analyzeSentiment("Yeah, this was a great soccer game of the German team!")
convertToBinaryResponse(sentiment)$SentimentGI
#> [1] positive
#> Levels: negative positive

Small example

The following demonstrates some of the functionality provided by SentimentAnalysis. It also shows its visualization and evaluation capabilities.

# Create a vector of strings
documents <- c("Wow, I really like the new light sabers!",
               "That book was excellent.",
               "R is a fantastic language.",
               "The service in this restaurant was miserable.",
               "This is neither positive or negative.",
               "The waiter forget about my a dessert -- what a poor service!")
 
# Analyze sentiment
sentiment <- analyzeSentiment(documents)
 
# Extract dictionary-based sentiment according to the QDAP dictionary
sentiment$SentimentQDAP
#> [1]  0.3333333  0.5000000  0.5000000 -0.3333333  0.0000000 -0.4000000
 
# View sentiment direction (i.e. positive, neutral and negative)
convertToDirection(sentiment$SentimentQDAP)
#> [1] positive positive positive negative neutral  negative
#> Levels: negative neutral positive
 
response <- c(+1, +1, +1, -1, 0, -1)
 
compareToResponse(sentiment, response)
#> Warning in cor(sentiment, response): the standard deviation is zero
#> Warning in cor(x, y): the standard deviation is zero
 
#> Warning in cor(x, y): the standard deviation is zero
#> Warning in cor(sentiment, response): the standard deviation is zero
#>                              WordCount SentimentGI NegativityGI
#> cor                        -0.18569534   0.9900115  -0.99748901
#> cor.t.statistic            -0.37796447  14.0440465 -28.16913204
#> cor.p.value                -0.37796447  14.0440465 -28.16913204
#> lm.t.value                 -0.37796447  14.0440465 -28.16913204
#> r.squared                   0.03448276   0.9801228   0.99498433
#> RMSE                        3.82970843   0.4501029   1.18665418
#> MAE                         3.33333333   0.4000000   1.10000000
#> Accuracy                    0.66666667   1.0000000   0.66666667
#> Precision                          NaN   1.0000000          NaN
#> Sensitivity                 0.00000000   1.0000000   0.00000000
#> Specificity                 1.00000000   1.0000000   1.00000000
#> F1                          0.00000000   0.5000000   0.00000000
#> BalancedAccuracy            0.50000000   1.0000000   0.50000000
#> avg.sentiment.pos.response  3.25000000   0.3333333   0.08333333
#> avg.sentiment.neg.response  4.00000000  -0.6333333   0.63333333
#>                            PositivityGI SentimentHE NegativityHE
#> cor                           0.9429542   0.4152274 -0.083045480
#> cor.t.statistic               5.6647055   0.9128709 -0.166666667
#> cor.p.value                   5.6647055   0.9128709 -0.166666667
#> lm.t.value                    5.6647055   0.9128709 -0.166666667
#> r.squared                     0.8891626   0.1724138  0.006896552
#> RMSE                          0.7136240   0.8416254  0.922958207
#> MAE                           0.6666667   0.7500000  0.888888889
#> Accuracy                      0.6666667   0.6666667  0.666666667
#> Precision                           NaN         NaN          NaN
#> Sensitivity                   0.0000000   0.0000000  0.000000000
#> Specificity                   1.0000000   1.0000000  1.000000000
#> F1                            0.0000000   0.0000000  0.000000000
#> BalancedAccuracy              0.5000000   0.5000000  0.500000000
#> avg.sentiment.pos.response    0.4166667   0.1250000  0.083333333
#> avg.sentiment.neg.response    0.0000000   0.0000000  0.000000000
#>                            PositivityHE SentimentLM NegativityLM
#> cor                           0.3315938   0.7370455  -0.40804713
#> cor.t.statistic               0.7029595   2.1811142  -0.89389841
#> cor.p.value                   0.7029595   2.1811142  -0.89389841
#> lm.t.value                    0.7029595   2.1811142  -0.89389841
#> r.squared                     0.1099545   0.5432361   0.16650246
#> RMSE                          0.8525561   0.7234178   0.96186547
#> MAE                           0.8055556   0.6333333   0.92222222
#> Accuracy                      0.6666667   0.8333333   0.66666667
#> Precision                           NaN   1.0000000          NaN
#> Sensitivity                   0.0000000   0.5000000   0.00000000
#> Specificity                   1.0000000   1.0000000   1.00000000
#> F1                            0.0000000   0.3333333   0.00000000
#> BalancedAccuracy              0.5000000   0.7500000   0.50000000
#> avg.sentiment.pos.response    0.2083333   0.2500000   0.08333333
#> avg.sentiment.neg.response    0.0000000  -0.1000000   0.10000000
#>                            PositivityLM RatioUncertaintyLM SentimentQDAP
#> cor                           0.6305283                 NA     0.9865356
#> cor.t.statistic               1.6247248                 NA    12.0642877
#> cor.p.value                   1.6247248                 NA    12.0642877
#> lm.t.value                    1.6247248                 NA    12.0642877
#> r.squared                     0.3975659                 NA     0.9732526
#> RMSE                          0.7757911          0.9128709     0.5398902
#> MAE                           0.7222222          0.8333333     0.4888889
#> Accuracy                      0.6666667          0.6666667     1.0000000
#> Precision                           NaN                NaN     1.0000000
#> Sensitivity                   0.0000000          0.0000000     1.0000000
#> Specificity                   1.0000000          1.0000000     1.0000000
#> F1                            0.0000000          0.0000000     0.5000000
#> BalancedAccuracy              0.5000000          0.5000000     1.0000000
#> avg.sentiment.pos.response    0.3333333          0.0000000     0.3333333
#> avg.sentiment.neg.response    0.0000000          0.0000000    -0.3666667
#>                            NegativityQDAP PositivityQDAP
#> cor                           -0.94433955      0.9429542
#> cor.t.statistic               -5.74114834      5.6647055
#> cor.p.value                   -5.74114834      5.6647055
#> lm.t.value                    -5.74114834      5.6647055
#> r.squared                      0.89177719      0.8891626
#> RMSE                           1.06840137      0.7136240
#> MAE                            1.01111111      0.6666667
#> Accuracy                       0.66666667      0.6666667
#> Precision                             NaN            NaN
#> Sensitivity                    0.00000000      0.0000000
#> Specificity                    1.00000000      1.0000000
#> F1                             0.00000000      0.0000000
#> BalancedAccuracy               0.50000000      0.5000000
#> avg.sentiment.pos.response     0.08333333      0.4166667
#> avg.sentiment.neg.response     0.36666667      0.0000000
 
# Optional visualization: plotSentimentResponse(sentiment$SentimentQDAP, response)

Dictionary generation

Research in finance and social sciences nowadays utilizes content analysis to understand human decisions in the face of textual materials. While content analysis has received great traction lately, the available tools are not yet living up to the needs of researchers. This package implements a novel approach named "**dictionary generation" to study tone, sentiment and reception of textual materials.

The approach utilizes LASSO regularization to extract words from documents that statistically feature a positive and negative polarity. This immediately reveals manifold implications for practitioners, finance research and social sciences: researchers can use R to extract text components that are relevant for readers and test their hypothesis based on these.

  • Proellochs, Feuerriegel and Neumann (2015): Generating Domain-Specific Dictionaries Using Bayesian Learning, Proceedings of the 23rd European Conference on Information Systems (ECIS 2015), Muenster, Germany. DOI: 10.2139/ssrn.2522884

License

SentimentAnalysis is released under the MIT License

Copyright (c) 2017 Stefan Feuerriegel & Nicolas Pröllochs

News

SentimentAnalysis 1.3-0

Version 1.3-0

New features and functions

  • Added new function extractWords() to retrieve all words from dictionary
  • Added new models for ridge regression, ordinary least squares and generalized linear model
  • Added flag to prefilter terms for dictionary generation

Changes

  • Fixed error in ruleLinearModel

Version 1.2-0

Changes

  • Minor fixes for CRAN release

Version 1.1-0

New features and functions

  • Added new function countWords() to count words without default preprocessing rules

Reference manual

It appears you don't have a PDF plugin for this browser. You can click here to download the reference manual.

install.packages("SentimentAnalysis")

1.3-2 by Stefan Feuerriegel, 2 months ago


https://github.com/sfeuerriegel/SentimentAnalysis


Report a bug at https://github.com/sfeuerriegel/SentimentAnalysis/issues


Browse source code at https://github.com/cran/SentimentAnalysis


Authors: Stefan Feuerriegel [aut, cre], Nicolas Proellochs [aut]


Documentation:   PDF Manual  


MIT + file LICENSE license


Imports tm, qdapDictionaries, ngramrr, moments, stringdist, SnowballC, XML, glmnet, spikeslab, ggplot2, mgcv

Suggests testthat, knitr, rmarkdown


See at CRAN