{"id":8,"date":"2019-09-03T09:15:50","date_gmt":"2019-09-03T09:15:50","guid":{"rendered":"http:\/\/nlp.pef.mendelu.cz\/?page_id=8"},"modified":"2025-03-14T13:40:22","modified_gmt":"2025-03-14T13:40:22","slug":"home","status":"publish","type":"page","link":"https:\/\/nlp.pef.mendelu.cz\/","title":{"rendered":"Home"},"content":{"rendered":"\n

English version<\/a><\/p>\n\n\n\n

\n
\n
\n

Natural Language Processing Group<\/h2>\n<\/div><\/div>\n\n\n\n
<\/div>\n\n\n\n
\n
\"\"<\/figure><\/div><\/div><\/div>\n\n\n\n
<\/div>\n\n\n\n
\n

Gathering and revealing the information<\/p>\n<\/div><\/div>\n<\/div>\n<\/div>\n\n\n\n

\n
\n
<\/div>\n\n\n\n
\n
\n
\n

Kdo jsme? <\/h2>\n<\/div><\/div>\n\n\n\n
\n
\n
\n

Jsme pracovn\u00ed skupina v\u011bnuj\u00edc\u00ed se problematice zpracov\u00e1n\u00ed p\u0159irozen\u00e9ho jazyka. Zam\u011b\u0159ujeme se na s\u00e9mantick\u00e9 zpracov\u00e1n\u00ed elektronick\u00fdch textov\u00fdch dat vytvo\u0159en\u00fdch v p\u0159irozen\u00fdch jazyc\u00edch s c\u00edlem odhalit znalosti skryt\u00e9 v t\u011bchto datech. Pou\u017e\u00edv\u00e1me zejm\u00e9na metody strojov\u00e9ho u\u010den\u00ed. Vyv\u00edj\u00edme webov\u00e9 i mobiln\u00ed aplikace, kter\u00e9 vyu\u017e\u00edvaj\u00ed algoritmy texminingu, pro \u0159e\u0161en\u00ed probl\u00e9m\u016f b\u011b\u017en\u00fdch u\u017eivatel\u016f a z\u00e1kazn\u00edk\u016f. Na\u0161e \u0159e\u0161en\u00ed jsou vyu\u017e\u00edv\u00e1na i komer\u010dn\u00edmi spole\u010dnostmi.<\/p>\n<\/div><\/div>\n<\/div>\n\n\n\n

\n
\n
\"\"<\/figure><\/div>\n\n\n

<\/p>\n<\/div><\/div>\n<\/div>\n<\/div>\n<\/div><\/div>\n<\/div>\n\n\n\n

<\/div>\n<\/div>\n\n\n\n
\n
<\/div>\n\n\n\n
\n
\n
\n

V\u00fdstupy<\/h2>\n<\/div><\/div>\n\n\n\n
\n
\n
\"\"<\/a>
Summary of Financial Articles<\/strong>

<\/figcaption><\/figure><\/div><\/div>\n\n\n\n
\n
\"\"<\/a>
N\u00e1stroje a aplikace pro zpracov\u00e1n\u00ed p\u0159irozen\u00e9ho jazyka<\/strong><\/figcaption><\/figure><\/div><\/div>\n\n\n\n
\n
\"\"<\/a>
Vyhled\u00e1v\u00e1n\u00ed odborn\u00edk\u016f<\/strong>
<\/figcaption><\/figure><\/div><\/div>\n\n\n\n
\n
\"\"<\/a>
Mobiln\u00ed aplikace Opinio<\/strong>
<\/figcaption><\/figure><\/div><\/div>\n<\/div>\n<\/div><\/div>\n<\/div>\n\n\n\n
<\/div>\n<\/div>\n\n\n\n
\n
<\/div>\n\n\n\n
\n
\n

U\u017eite\u010dn\u00e9<\/h2>\n\n\n\n

P\u0159i na\u0161\u00ed pr\u00e1ci vyu\u017e\u00edv\u00e1me r\u016fzn\u00e9 n\u00e1stroje pro p\u0159\u00edpravu a n\u00e1slednou anal\u00fdzu textov\u00fdch dokument\u016f. Jde nap\u0159\u00edklad o lemmatizaci<\/strong>, stemming<\/strong>, odstran\u011bn\u00ed stopslov<\/strong>, identifikace sentimentu<\/strong>, a dal\u0161\u00ed. Najdete zde i na\u0161eho chatbota<\/strong>, kter\u00fd doporu\u010duje mobiln\u00ed telefony.<\/p>\n<\/div><\/div>\n<\/div>\n\n\n\n

<\/div>\n<\/div>\n\n\n\n
\n
<\/div>\n\n\n\n
\n
\n

Zpracov\u00e1n\u00ed p\u0159irozen\u00e9ho jazyka<\/h2>\n\n\n\n

Odhaduje se, \u017ee p\u0159es 80 % dat je dnes ulo\u017eeno v podob\u011b text\u016f (novinov\u00e9 \u010dl\u00e1nky, e-maily, blogy, p\u0159\u00edsp\u011bvky na Facebooku atd.), je\u017e jsou m\u00e1lo nebo v\u016fbec strukturovan\u00e9. Pot\u0159eba anal\u00fdzy textov\u00fdch dat v sou\u010dasn\u00e9 dob\u011b roste a st\u00e1v\u00e1 se velmi komer\u010dn\u011b zaj\u00edmavou oblast\u00ed. C\u00edlem analytick\u00fdch \u00faloh je odhalen\u00ed p\u0159edem nezn\u00e1m\u00fdch znalost\u00ed obsa\u017een\u00fdch v t\u011bchto datech netrivi\u00e1ln\u00edmi metodami. V\u00fdsledky nach\u00e1zej\u00ed uplatn\u011bn\u00ed v oblastech marketingu, po\u010d\u00edta\u010dov\u00e9 bezpe\u010dnosti, informa\u010dn\u00edch slu\u017eeb, liter\u00e1rn\u00edch re\u0161er\u0161\u00ed, \u0159\u00edzen\u00ed lidsk\u00fdch zdroj\u016f, boje proti terorismu apod.<\/p>\n\n\n\n

\n
\n
\n

Pr\u00e1ce s textov\u00fdmi daty je obecn\u011b velmi obt\u00ed\u017en\u00e1. Data jsou v\u011bt\u0161inou nestrukturovan\u00e1 a maj\u00ed zcela jin\u00fd charakter ne\u017e data \u010d\u00edseln\u00e1 (slo\u017eit\u00e1 gramatika, r\u016fzn\u00fd v\u00fdznam slov, subjektivita, ironie apod.). Postupy uspokojiv\u011b funguj\u00edc\u00ed pro jednu dom\u00e9nu nemusej\u00ed nav\u00edc fungovat pro dom\u00e9nu jinou. Snahou je nasadit metody z oblasti Data Miningu. Tato zral\u00e1 a dob\u0159e rozvinut\u00e1 discipl\u00edna se tak\u00e9 zam\u011b\u0159uje se na hled\u00e1n\u00ed skryt\u00fdch znalost\u00ed v datech, ale pracuje s vysoce strukturovan\u00fdmi \u010d\u00edseln\u00fdmi daty. Je tedy v\u00fdhodn\u00e9 textov\u00e1 data p\u0159ipravit tak, aby na n\u011b byly metody Data Miningu pou\u017eiteln\u00e9. To vy\u017eaduje nasazen\u00ed postup\u016f z oblast\u00ed, jak\u00fdmi jsou zpracov\u00e1n\u00ed p\u0159irozen\u00e9ho jazyka, statistika, strojov\u00e9 u\u010den\u00ed, lingvistika a dal\u0161\u00edch.<\/p>\n<\/div><\/div>\n<\/div>\n\n\n\n

\n
\n
\"\"<\/figure><\/div><\/div><\/div>\n<\/div>\n<\/div>\n\n\n\n
\n
\n
\n
\"\"<\/figure><\/div><\/div><\/div>\n<\/div>\n\n\n\n
\n
\n

Pro anal\u00fdzu textov\u00fdch dat jsou v r\u00e1mci v\u00fdzkumu aplikov\u00e1ny zejm\u00e9na metody strojov\u00e9ho u\u010den\u00ed s u\u010ditelem (klasifikace), bez u\u010ditele (shlukov\u00e1n\u00ed, v\u00fdb\u011br atribut\u016f, hled\u00e1n\u00ed asociac\u00ed) a \u010d\u00e1ste\u010dn\u011b \u0159\u00edzen\u00e9ho u\u010den\u00ed (semi-supervised) a jejich kombinace. C\u00edle v\u00fdzkumu zahrnuj\u00ed kategorizaci textov\u00fdch dokument\u016f, vyhled\u00e1v\u00e1n\u00ed dokument\u016f na z\u00e1klad\u011b podobnosti, odhalov\u00e1n\u00ed s\u00e9mantiky skupin dokument\u016f, hled\u00e1n\u00ed atribut\u016f, kter\u00e9 vyjad\u0159uj\u00ed v\u00fdznam, anal\u00fdzu sentimentu a dal\u0161\u00ed. V \u00favahu se berou specifick\u00e9 vlastnosti a omezen\u00ed t\u011bchto \u00faloh, jako nap\u0159\u00edklad mal\u00fd po\u010det vhodn\u00fdch p\u0159\u00edklad\u016f, obrovsk\u00e9 objemy a dimenzionalita dat, \u0159\u00eddkost vektor\u016f reprezentuj\u00edc\u00edch data, nevyv\u00e1\u017een\u00fdch t\u0159\u00edd \u010di mnohojazy\u010dnost. Tyto zvl\u00e1\u0161tnosti jsou typick\u00e9 pro soubory dat vytv\u00e1\u0159en\u00fdch samotn\u00fdmi u\u017eivateli na soci\u00e1ln\u00edch s\u00edt\u00edch, mikrobloz\u00edch nebo diskusn\u00edch f\u00f3rech (na rozd\u00edl od v\u011bdeck\u00fdch prac\u00ed \u010di novinov\u00fdch \u010dl\u00e1nk\u016f). Budouc\u00ed v\u00fdzkum se zam\u011b\u0159\u00ed tak\u00e9 na anal\u00fdzu vztah\u016f mezi textov\u00fdmi daty (zpr\u00e1vy, ekonomick\u00e9 souhrny, p\u0159\u00edsp\u011bvky a liky na soci\u00e1ln\u00edch s\u00edt\u00edch) a r\u016fzn\u00fdmi ekonomick\u00fdmi jevy, jako je nap\u0159\u00edklad v\u00fdvoj cen akci\u00ed.<\/p>\n<\/div><\/div>\n<\/div>\n<\/div>\n\n\n\n

\n
\n
\n

P\u0159i realizaci v\u00fdzkumu jsou vyv\u00edjeny a nasazov\u00e1ny n\u00e1stroje pro p\u0159edzpracov\u00e1n\u00ed dat, zahrnuj\u00edc\u00edch stemming, odstran\u011bn\u00ed stop slov, ur\u010den\u00ed slovn\u00edch druh\u016f, kontrolu pravopisu a dal\u0161\u00ed. Pro transformaci dat do form\u00e1tu vhodn\u00e9ho pro algoritmy strojov\u00e9ho u\u010den\u00ed a nejr\u016fzn\u011bj\u0161\u00ed softwarov\u00e9 n\u00e1stroje je pr\u016fb\u011b\u017en\u011b vyv\u00edjena unik\u00e1tn\u00ed aplikace s voliteln\u00fdm grafick\u00fdm u\u017eivatelsk\u00fdm rozhran\u00edm. V\u00fdzkum zahrnuje tak\u00e9 aplikaci profesion\u00e1ln\u00edho komer\u010dn\u00edho \u010di open-source programov\u00e9ho vybaven\u00ed (C5, Cluto, Weka, IBM SPSS Modeler a dal\u0161\u00ed).<\/p>\n<\/div><\/div>\n<\/div>\n\n\n\n

\n
\n
\"\"<\/figure><\/div><\/div><\/div>\n<\/div>\n<\/div>\n<\/div><\/div>\n<\/div>\n\n\n\n
<\/div>\n<\/div>\n<\/div><\/div>\n","protected":false},"excerpt":{"rendered":"

English version Natural Language Processing Group Gathering and revealing the information Kdo jsme? Jsme pracovn\u00ed skupina v\u011bnuj\u00edc\u00ed se problematice zpracov\u00e1n\u00ed […]<\/p>\n","protected":false},"author":1,"featured_media":0,"parent":0,"menu_order":0,"comment_status":"closed","ping_status":"closed","template":"","meta":{"footnotes":""},"class_list":["post-8","page","type-page","status-publish","hentry"],"_links":{"self":[{"href":"https:\/\/nlp.pef.mendelu.cz\/index.php\/wp-json\/wp\/v2\/pages\/8","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/nlp.pef.mendelu.cz\/index.php\/wp-json\/wp\/v2\/pages"}],"about":[{"href":"https:\/\/nlp.pef.mendelu.cz\/index.php\/wp-json\/wp\/v2\/types\/page"}],"author":[{"embeddable":true,"href":"https:\/\/nlp.pef.mendelu.cz\/index.php\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/nlp.pef.mendelu.cz\/index.php\/wp-json\/wp\/v2\/comments?post=8"}],"version-history":[{"count":77,"href":"https:\/\/nlp.pef.mendelu.cz\/index.php\/wp-json\/wp\/v2\/pages\/8\/revisions"}],"predecessor-version":[{"id":668,"href":"https:\/\/nlp.pef.mendelu.cz\/index.php\/wp-json\/wp\/v2\/pages\/8\/revisions\/668"}],"wp:attachment":[{"href":"https:\/\/nlp.pef.mendelu.cz\/index.php\/wp-json\/wp\/v2\/media?parent=8"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}