Hi Chris, Nice work generally. I love this kind of thing hence why I love GATE which uses both Stanford toolkits and Tika as well :)
On Fri, Nov 20, 2015 at 8:37 AM, <[email protected]> wrote: > > Thamme and I added a wiki page for Tika/Stanford NER and Apache > OpenNLP integration: > > http://wiki.apache.org/tika/TikaAndNER > > Some interesting results... particularly on the CoreNLP NER results shown in the wiki page. As I've noticed before, lots of false positives within organization recognition. Looks like a rule based approach would align these. NER_DATE: 2009NER_DATE: 1963NER_DATE: 1663NER_DATE: 1982NER_DATE: 1979NER_LOCATION: GonvilleNER_LOCATION: Einstein <<< Not a locationNER_LOCATION: LondonNER_LOCATION: CambridgeNER_LOCATION: Santa CruzNER_ORGANIZATION: Leiden UniversityNER_ORGANIZATION: NASANER_ORGANIZATION: CBENER_ORGANIZATION: Brief History of Time <<< Not an OrganizationNER_ORGANIZATION: University of CaliforniaNER_ORGANIZATION: Cambridge Lectures Publications Books Images Films <<< Books Images Films in not in Organization nameNER_ORGANIZATION: Caius CollegeNER_ORGANIZATION: Royal Society <<< Incomplete Organization nameNER_ORGANIZATION: About Stephen The Computer Stephen <<< Not an organizationNER_ORGANIZATION: US National Academy of ScienceNER_ORGANIZATION: Department of Applied MathematicsNER_ORGANIZATION: ESANER_ORGANIZATION: The Universe <<< Not an OrganizationNER_ORGANIZATION: Sally Tsui Wong-Avery Director of Research <<< Not an OrganizationNER_ORGANIZATION: the University of CambridgeNER_ORGANIZATION: Theoretical Physics <<< Not an OrganizationNER_ORGANIZATION: Baby Universe <<< Not an OrganizationNER_PERSON: EinsteinNER_PERSON: P. OeschNER_PERSON: R. BouwensNER_PERSON: GeorgeNER_PERSON: Stephen HawkingNER_PERSON: Isaac NewtonNER_PERSON: D. MageeNER_PERSON: AnnieNER_PERSON: G. IllingworthNER_PERSON: StephenNER_PERSON: Dennis Stanton Avery Wiki page is really helpful to understand extent and functionality available within Tika now. Excellent work folks. Lewis
