To add to Tim's clarification. In addition, this enables you (or anyone for that matter) to implement your /their own method for these types.
--Guergana -----Original Message----- From: Miller, Timothy [mailto:timothy.mil...@childrens.harvard.edu] Sent: Thursday, April 30, 2020 7:53 AM To: dev@ctakes.apache.org Subject: Re: how to activate inactive features in cTAKES? [EXTERNAL] [SUSPICIOUS] * External Email - Caution * Akram, the typesystem in ctakes was created by a project with the aim of specifying things that are useful, without specifying implementations for them all. There are many items in the data model that there are no ctakes modules to fill. The idea was that when people bring things online there are placeholders for that information, so that new functionality is not added in a completely ad hoc way. So of the examples you describe: - discoveryTechnique is always the same because you are running the same pipeline - confidence is not filled in by the dictionary lookup -- the current method used does not generate a confidence score - disambiguated is not filled but is technically correct because there is no disambiguation algorithm running - polarity, uncertainty, conditional, generic, historyOf, can be filled in by certain pipelines. You will have to add them after the DictionarySubPIpe to see them filled in. Tim ________________________________________ From: Akram <as...@yahoo.com.INVALID> Sent: Thursday, April 30, 2020 4:37 AM To: dev@ctakes.apache.org Subject: how to activate inactive features in cTAKES? [EXTERNAL] * External Email - Caution * Hi I can extract many tags when I use the default .piper in cTakes Tags such as LabMention, AnatomicalSiteMention, ProcedureMention, etc they all extracted from applying this piper ---------------------------------------------------------------- load DefaultTokenizerPipeline load DictionarySubPipe writeHtml writeXmis ---------------------------------------------------------------- The problem is there are some features that do not change no matter the text change. most importantly confidence which is always 0 How can I get the confidence of each term? other features such discoveryTechnique is always 1 polarity always 0 uncertainty always 0 conditional always false generic always false historyOf always 0 score always 0 disambiguated always flase how can I get these features working and where can I find more info about these features and what do they mean? Thanks
