This container adds sentiment annotations to the data. Since context matters a lot for choosing the right dictionary, we do not provide dictionaries within the framework. It is up to the user to choose a dictionary - either download a dictionary from one of the many sources (some sources are given below) or provide an own dictionary. Please note that dictionaries do not necessarily need to indicate positive or negative valences, you can also apply dictionaries promoting two categories.
Users need to store the dictionary as a comma-separated file (csv) in the folder 00_custom/sentiment_dictionaries. The csv-file needs to have two columns only - a column for words (word) and a column for the sentiment value (value), with the variable names on the first row, e.g.
word,value
abandon,-1
abandoned,-1
If users provide multiple dictionaries, please indicate in the settings which dictionary you want to choose. If no dictionary is set, the system automatically takes the first one in the (alphabetic) list of files - irrespective of the language.
Some sources for dictionaries:
R based functions.
Valentin Gold