<?xml version="1.0"?>
<feed xmlns="http://www.w3.org/2005/Atom" xml:lang="en">
	<id>https://subvention.zooid.org/index.php?action=history&amp;feed=atom&amp;title=Corporation_de_d%C3%A9veloppement_communautaire_du_Granit</id>
	<title>Corporation de développement communautaire du Granit - Revision history</title>
	<link rel="self" type="application/atom+xml" href="https://subvention.zooid.org/index.php?action=history&amp;feed=atom&amp;title=Corporation_de_d%C3%A9veloppement_communautaire_du_Granit"/>
	<link rel="alternate" type="text/html" href="https://subvention.zooid.org/index.php?title=Corporation_de_d%C3%A9veloppement_communautaire_du_Granit&amp;action=history"/>
	<updated>2026-10-22T00:34:29Z</updated>
	<subtitle>Revision history for this page on the wiki</subtitle>
	<generator>MediaWiki 1.39.11</generator>
	<entry>
		<id>https://subvention.zooid.org/index.php?title=Corporation_de_d%C3%A9veloppement_communautaire_du_Granit&amp;diff=3560&amp;oldid=prev</id>
		<title>DavidM: Created page with &quot;{{Subventions |import ref=Subventions 2215 |Code postal=G6B 1X8 |Rue=4982, rue Champlain, bureau 105 |Ville=Lac-Mégantic |Ministère=MESSF(SACA) }}&quot;</title>
		<link rel="alternate" type="text/html" href="https://subvention.zooid.org/index.php?title=Corporation_de_d%C3%A9veloppement_communautaire_du_Granit&amp;diff=3560&amp;oldid=prev"/>
		<updated>2011-01-29T12:06:54Z</updated>

		<summary type="html">&lt;p&gt;Created page with &amp;quot;{{Subventions |import ref=Subventions 2215 |Code postal=G6B 1X8 |Rue=4982, rue Champlain, bureau 105 |Ville=Lac-Mégantic |Ministère=MESSF(SACA) }}&amp;quot;&lt;/p&gt;
&lt;p&gt;&lt;b&gt;New page&lt;/b&gt;&lt;/p&gt;&lt;div&gt;A full week of learning [http://en.wikipedia.org/wiki/General_Architecture_for_Text_Engineering GATE] text mining/information extraction language processing and talks. [http://gate.ac.uk/wiki/TrainingCourseAug2010/ Session wiki]&lt;br /&gt;
&lt;br /&gt;
[[File:GATE_screenshot.png|900px|GATE developer screenshot]]&lt;br /&gt;
&lt;br /&gt;
GATE is written in Java and very Java centric. This makes it portable, fast, and heavyweight. A programming library is available.  It&amp;#039;s 14 years old and has many users and contributors.&lt;br /&gt;
&lt;br /&gt;
= Using GATE developer =&lt;br /&gt;
&lt;br /&gt;
* GATE developer is used to process sets of Language Resources in Corpus using Processing Resources. They are typically saved to a serialized Datastore.&lt;br /&gt;
* ANNIE, VG (verb group) processors.&lt;br /&gt;
* Preserve formatting embeds tags in HTML or XML.&lt;br /&gt;
** Different strengths using GATE&amp;#039;s graph (node/offset) based XML vs. preserved formatting (original xml/html)&lt;br /&gt;
&lt;br /&gt;
= Information Extraction =&lt;br /&gt;
&lt;br /&gt;
* IR - retrieve docs&lt;br /&gt;
* IE - retreive structured data&lt;br /&gt;
&lt;br /&gt;
* Knowledge Engineering - rule based&lt;br /&gt;
* Learning Systems - statistical&lt;br /&gt;
&lt;br /&gt;
Old Bailey IE project  - old english (Online)&lt;br /&gt;
&lt;br /&gt;
* POS - assigned in Token (noun, verb, etc)&lt;br /&gt;
&lt;br /&gt;
* Gazateer - gotcha, have to set initialization parameter listsURL before it&amp;#039;s loaded. Must also &amp;quot;save and reinitialize.&amp;quot;&lt;br /&gt;
* Gazeteer creates Lookups, then transducer creaties named entities&lt;br /&gt;
* Then orthomatcher (spelling features in common) coreference associates those&lt;br /&gt;
&lt;br /&gt;
* Annotation Key sets and annotation comparing&lt;br /&gt;
** Need setToKeep key in Document Reset for any pre annotated texts&lt;br /&gt;
&lt;br /&gt;
== Evaluation / Metrics ==&lt;br /&gt;
&lt;br /&gt;
* Evaluation metric - mathematically against human annotated&lt;br /&gt;
* Scoring - performance measures for annotation types&lt;br /&gt;
&lt;br /&gt;
* Precision = correct / correct + spurious&lt;br /&gt;
* Recall = correct / correct + missing&lt;br /&gt;
* F-measure is precision and recall (harmonic mean)&lt;br /&gt;
* F=2⋅(precision⋅recall / precision+recall)&lt;br /&gt;
* GATE supports average, strict, lenient&lt;br /&gt;
&lt;br /&gt;
* Result types - Correct, missing, spurious, partially correct (overlapped)&lt;br /&gt;
&lt;br /&gt;
* Tools &amp;gt; Annotations Diff - comparing human vs machine annotation&lt;br /&gt;
&lt;br /&gt;
* Corpus &amp;gt; Corpus quality assurance - compare by type&lt;br /&gt;
* (B has to be the generated set)&lt;br /&gt;
&lt;br /&gt;
* Annotation set transfer (in tools) - transfer between docs in pipeline&lt;br /&gt;
** useful for eg HTML that has boilerplate&lt;br /&gt;
&lt;br /&gt;
= Other notes =&lt;br /&gt;
&lt;br /&gt;
== Lucene data store and ANIC ==&lt;br /&gt;
&lt;br /&gt;
* Use &amp;lt;null&amp;gt; for default set&lt;br /&gt;
* Go to Datastore for queries&lt;br /&gt;
** eg {Person}({Token})+{Money}&lt;br /&gt;
* Useful for debugging JAPE and results&lt;br /&gt;
&lt;br /&gt;
[[File:GATE-lucene-person-money.png|800px]]&lt;br /&gt;
&lt;br /&gt;
== To investigate ==&lt;br /&gt;
&lt;br /&gt;
* markupAware for HTML/XML (keeps tags in editor)&lt;br /&gt;
* AnnotationStack&lt;br /&gt;
* Advanced Options&lt;br /&gt;
&lt;br /&gt;
= JAPE =&lt;br /&gt;
&lt;br /&gt;
* Rules based on tokens and lookups&lt;br /&gt;
&lt;br /&gt;
== To review, gotchas ==&lt;br /&gt;
&lt;br /&gt;
* Rule types : first takes only first match, excludes compound&lt;br /&gt;
** a? b for &amp;quot;a b&amp;quot; will match &amp;quot;a b&amp;quot; &lt;br /&gt;
* multiplexor tranducers&lt;br /&gt;
* multi-constraint statements&lt;br /&gt;
* macros&lt;br /&gt;
* To reuse created annotations has to be a separate rule&lt;br /&gt;
&lt;br /&gt;
= To follow up =&lt;br /&gt;
&lt;br /&gt;
* WebSphinx crawler CREOLE plugin&lt;br /&gt;
&lt;br /&gt;
= Demos =&lt;br /&gt;
&lt;br /&gt;
* Mímir for querying large volumes of data (uses MG4J)&lt;br /&gt;
* Translating parts of speech between languages using Compound editor and Alignment editor&lt;br /&gt;
* Predicate extractor (MultiPaX)&lt;br /&gt;
** Mixed results at best&lt;br /&gt;
* OwlExporter&lt;br /&gt;
** NLP ontology&lt;br /&gt;
&lt;br /&gt;
= Conclusions =&lt;br /&gt;
&lt;br /&gt;
While it can do a lot out of the box and benefits from development time and breadth of connectivity, to be useful to more than patient specialists, it needs usability testing. A lot of things are inobvious and too domain specific that with a bit of work could be more broadly useful. Interaction could include a lot more immediate, useful and interesting looking displays. A web based version could have these features. However the team seems somewhat ambivalent about development. :)&lt;br /&gt;
&lt;br /&gt;
{{Blikied|Aug 30, 2010}}&lt;br /&gt;
&lt;br /&gt;
[[Category:SemWeb]]&lt;/div&gt;</summary>
		<author><name>DavidM</name></author>
	</entry>
</feed>