# Dutch confusion sets
# Line format:
# <word1>|<description1>; <word2>|<description2>; <factor>   # optional comment
#   <word1> and <word2> are words that can easily be confused
#   <description> will be used in the error message to explain the word (optional)
#   <factor> is the factor of how much more the other word must be more
#            probable so the text is considered potentially incorrect.
#            Use a higher value for better precision but lower recall.
#   Precision (p) and recall (r) values in the comments come from ConfusionRuleEvaluator
#   The number after recall is the number of sentences used for evaluation.
# Order is relevant for ambiguous cases like 'know' ('no' or 'now') where the match
# is used whose pair comes first in this file.
# 
bereiden|klaarmaken; berijden|rijden op; 1000000;       # p=1.000, r=0.252, 648+106, 3grams, 2019-01-02
boord|kraag; bord|plank, etensbord; 1000000;       # p=1.000, r=0.365, 979+929, 3grams, 2019-01-02