0
$\begingroup$

I am trying to perform string delivery classification (X: delivery_string, y: delivery_string_relevance (values 0 or 1)). I am using DTM (document term matrix) for the feature extraction, and experimenting with text-mining algorithms. However, I still believe that I can preprocess my raw strings better (before creating the DTM) by extracting delivery dates and obtaining one common format representation for all of them. Any suggestions on good date parsers? datefinder.find_dates() in Python is doing quite a decent job, however still fails for some cases, such like when string contains "delivery time 10 - 12 days" it extracts it as 10-12-2017.

Sample data (language used Dutch): beschikbaar vanaf 04 okt // beschikbaar vanaf: 05-08-2017 // 8 - 9 weken // 8 - 9 werkdagen // 8 dag(en)

$\endgroup$
1
$\begingroup$

The best option is dateparser, whose goal is to parse natural language dates with support for Dutch. There is another package recurrent which does natural language parsing of dates focusing on recurring events.

| improve this answer | |
$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.