Skip to content

Latest commit

 

History

History
106 lines (60 loc) · 2.81 KB

README.md

File metadata and controls

106 lines (60 loc) · 2.81 KB
Raw

Sentida

Sentida is a Danish sentiment analysis tool. Sentida sentiment-scores sentences by the word, and gives either a mean sentiment score or a total score. Sentida uses sophisticated methods to take adverb-modifiers, exclamation marks and negations into account when sentiment-scoring.

Downloading Sentida for the first time

R

if(!require("devtools")) install.packages("devtools")

devtools::install_github("Guscode/Sentida")

library(Sentida)

Python

pip install sentida

Using Sentida

R

The sentida function takes in a string and an output argument. The output argument can be the default "total" or "mean". Output = "total" will provide an accumulated sentiment score for the string. Output = "mean" will provide a mean sentiment score per word in the string.

Python

The sentida function in the Sentida class takes four arguments: The input text, the desired output, normalization, and the desired speed / accuracy prioritization. "mean" and "total" return a single float value and "by_sentence_mean" and "by_sentence_total" return a list of length equal to the amount of sentences.

Speed can be set to fast for a much less accurate but 180% faster implementation.

from sentida import Sentida
Sentida().sentida(
                text,
                output = ["mean", "total", "by_sentence_mean", "by_sentence_total"],
                normal = True,
                speed = ["normal", "fast"]
                )

Examples

Sentida in a political context:

Total score:

sentida("Abort er mord", output = "total") -> -4.67

Mean score:

sentida("Abort er mord", output = "mean") -> -1.56

Effect of exclamation mark:

sentida("Abort er mord!", output = "total") -> -6.02

Sentida in a commercial context:

Total score:

sentida("Colgate er godt og smager dejligt", output = "total") -> 5

Effect of adverb modifiers:

sentida("Colgate er godt og smager mega dejligt", output = "total") -> 5.8

Effect of negations:

sentida("Colgate er ikke godt og smager ikke dejligt", output = "total") <- -4.34

Working with æøå

Working with æøå in R can cause certain problems. When workin with R in mac os, it is necessary to run the following code, in order to work with æøå.

Sys.setlocale(category = "LC_ALL", locale = "UTF-8")

Furthermore, encoding necessary files in UTF-8 format, will make æøå accessible to Rstudio.

Creators

Sentida has been created by three Cognitive Science bachelor students at Aarhus University:

Lars Kjartan Bacher Svendsen

Jacob Aarup Dalsgaard

Gustav Aarup Lauridsen

For commercial use, please contact: Email: [email protected], [email protected], [email protected]

The Python implementation has been created by:

Esben Kran

Søren Orm