Hi, Its an announcement for the 'official' release of GPoSTTL. GPoSTTL has been developed as an open-source alternative for TreeTagger, a Penn Treebank parts-of-speech tagger and a crucial component of Anubadok. With this availability Anubadok's non-GPL dependence is OPTIONAL from now. I should thank Progga for FreeBSD patches and Prof. Y. Someya for permitting the use of his lemma list.
GPoSTTL is available from "http://www.imsc.res.in/~golam/gposttl/" GPoSTTL is derived from Brill's tagger which is a rule-based tagger unlike TreeTagger which is a statistical tagger. Being open-source, it is possible to add new verbs like "blog" in its database quite easily, but that is not possible for TreeTagger. Naturally, TreeTagger's performance goes down while translating PO files as they contain many such new terms. So even though performance wise, TreeTagger is rated slightly higher than Brills tagger but in my experience (using with po_anubadok) difference isn't significant. Moreover, GPoSTTL is less than one tenth in size of TreeTagger. Recently, I have switched to GPoSTTL while using po_anubadok. Further, while digging through it, I realized that its possible to improve current tagging method of GPoSTTL. So any of you are interested in getting your hand dirty with NLP stuff then it could be a nice playground! Please mind that it is a parts-of-speech tagger for English. So having a smart open-source tagger would be greatly helpful for any machine translator that aims to translate from English to XX. Further, it could be a crucial component for futuristic Grammar correcting program for any English word processor. Please feel free to use, comment... Cheers, Golam -- http://www.imsc.res.in/~golam/ ------------------------------------------------------- Using Tomcat but need to do more? Need to support web services, security? Get stuff done quickly with pre-integrated technology to make your job easier Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronimo http://sel.as-us.falkag.net/sel?cmd=lnk&kid0709&bid&3057&dat1642 _______________________________________________ Bengalinux-core mailing list Bengalinux-core@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bengalinux-core