Text Preprocessing

[@khairulikhwanazman6094]

I need to save the clean and preprocessed word from orange in excel but whenever I save data it just revert back to the original data

[@eylmaz6696]

does orange have Cumulatif distribuation function and probability distribution function to get out the results ?

[@OrangeDataMining]

No sure what you wish to achieve, but these options are available in the Distributions widget.

[@eylmaz6696]

@@OrangeDataMining for clustering on kmeans algorithm ? which one is important silhouıette skor or centering on the intersection is important by checkng scatter polot?

[@eylmaz6696]

@@OrangeDataMining for kmeans clustering ; how can i make a comment for result ? for instance, relation between ; I have anxeity, I dont have anxiety ; I sleep much ; I dont sleep much. When I cluster them, will i comment it by using the silhouette score maximum ?

[@OrangeDataMining]

@@eylmaz6696 Apologies, I don't quite understand the question.

[@eylmaz6696]

@@OrangeDataMining do you have support mail or phone ? can i get one question

[@gabrielapinto5306]

I am finding it difficult to adapt all that to tweets written in Portuguese. Does orange have a solution?

[@OrangeDataMining]

Yes. The tokenizer remains the same. Stopwords are available for Portuguese, too. Same for lemmatization (only UDPipe). SBERT and FastText also support pt. In summary, most language specific methods support Portuguese (some also pt-br), others are language independent.

[@neilirvine7129]

Love it!

[@nadiamaelaniulfah1100]

does orange not support arabic? orange told "no text found" when i'm uploading my arabic corpus. any solution for this?🥲

[@OrangeDataMining]

Orange supports Arabic to some extent. There is Arabic lemmatizer with UDPipe, stopwords from NLTK, and embedders in Document Embedding. Your error tells you you are likely missing a text variable in your data. Please head to our discussions board (github.com/biolab/orange3-text/discussions) where we pinpoint your problem.