NLP BE CSE Yagai Naik Roll no. 32 Hssiqnment L Date Page 2 What is NLPZ Explain the diftexent stages i n NLP Natural Lanquage ocessinq is the' pYocess at computer analy.sis at input provided in a numan lan unae and convexsioD o this input into a usetul foxm yepresentaHan Natural lanquage pyocessinq 1s concexned wlfh Computahonal models ot the development aspects of human lanquage PYOcessing l h e tallowinq axe the tuwo main yeasons for uch developments Develop cutomatic tool tox natuxa longuoge processino- Caoin better undexstandinq o human communication, The input and output af the NLP is text OY Speech Stageg in NLP . Lexical Analysis Lexical Analusis is the tivst staqe in NLe T is also known as marphaloqical analusis At his stagethe stxuctuxe af the woYds isidentilied and analused.. Lexicon at a lanquage pmeans the collection ot wods anod phrases 1na language. -Lexical onalysis is divioling the whale poytion at text into paxoaraphs sentences ano words. Date Page 2 Suntactic Analysis lPaysing TH iovelves abalysis ot woxds in the sentence o aYamamay ond oxdevina wovdg a way that shows t h e Yelatiooship amona the words. Ihe sent ence such as he schoal aaes ginl is ejected by Enalish suntactic analuSey 3. Semantic Hnalusis Semantic atnalysis draws the exact meanin Oxthe dicionaru meanin he hoa the text text is ehecked fox meaninqtulness. i s dane by napping syntacticstruchuxes and abjects in the task domain Ihe semantic analuser nealects sentences Such ashot ice-cleam' 4 Dig couYceTntegranion Tbe rneaninq o anu sentence depends upon bet he meanina at the sentence ust e it. Fukthermovet also brins abod he meanina ot immediatelu ollowina sentence Fox example Meena is agi she qoes to scboal hexe she is a dependeneu poinhnq to Meena 5. Roamatic Analusis Durina this, what was said is Ye-inteYpYeted- on whot t tyuly meant t contains Date Page oevinq ot those aspechs o+ language which necessit.ate yeal woxld knowledae For example,John saw May in a gavdlen inth a cat', hexe we.cant say hat John is with cat o Maxus Lwith cot. 2.List and explain some impoytant opplicahons of NLP? Applicationg of NLP l Machine tyansloion I n nachine manslahon he tronslaion a the textin one buman lanauage to onothey buman lanauaaes is pertoxraed autnmatically. Foy Pextornnina he tran.slation t is impox-ant to bave kbowledae ot Hbe wovds and phrases, 1aamay o two lanauaqes hot aye invalved in ranslahon sermanhcs of he lanquaaes and he Knawledae o ae waYd 2peech ecoqniiar peech xecoQDihon is Hhe pYocess wbeye. the acoLushc Speech gianal.g axe nappedD he set at words As thexe is Lwide variation i n the pronounciation the elOYdbomanum and sea for exompleseoa acougheQmbiauiies ike in he Date Page NeSt and interest. 3. Speech sunthesis Automatic production o Spe ech is known as speecb sunthesis Enmeansspeakinq a sentence innatuYal lanqAoge IbespeecheuDthesis Sustem yeads raailon your telepbones_ax veads stoxubooks for you kor aeneyainq the uterance.g text PYOcessinq .s yeguiYed,so, NLP 1s an mpoxtant component in speech sunthesis Sugtem Intoymation xetyeval T toraation retrieval be xelevant docuaents velatedto he usey's Gueries axe identitiedIn Idoxaation veieval indexina quevu moditicohon, wrd sense disanbiquation, and knowledqe bases ave used o y eDhaneina the peroYmarnce Fox example , woronet and Lonaman Dictionary aContenmporaruy Enalish (LDoCE) Ave some usetul esauYCeG or Ifoymation yetyieval xeseqrch 5 Tntormation extyaction Itoxmaionextraation 1s one ot the mast sianiticont applicah.oDsaf NLP Tisused tar extracinq stxuehuxeo intornah'an Yam unstxuctuxedOr Gemisructured machine-xcadable documents Date Page nfoxmationexraction sustem captures and autputs actual infoxmahon contained within adocument Idoxronahon extrachom94steroa a s u b s e to d eniieg ettsnerteInformaction Awithina documenthat its t h e pxedetined emplate. 6 Question angwerina h e 9ues-tion answersSustem yies to hnd out h e coxyectansLuer or paxt Ot he text whexe he _answex appeaxs tox t h e given 9ue shon ano a s e t of docunments lhe 9uesion answerS Sgtem eurns a ull docum ent bat is yelevont to the. usexls queruTbe 2uesioh answey systerm usesan indormatiorn extracion system for idenifuinq he enitiesin the text. A queshion answexina sustem needed roave NLP hon an intoxmation xetxieval sustem ox an intoymaiob extYachon Sustem N needed process analysis ct 2ueshion and peYiong o ext and also semantic as well os backayound knowledqe tn answercertain ypea 1 gueshions extSummarizohon LextSumnavizahionpaeans cYeatoqshox coxectSumrmaxu a lonaex 4ext dacuments Autonnatiic extsumnarization assist S 1fh-aPpropiate in less hme Date Page - NLP has an important vole in developina an autormatic text summ.avization TextSuamaxization invalvessuntoacchic semanhcs a n d dis.couxGse leve processin.a o text. 8. 6entiment Analusis. Sentiment Analusis is also xetexyed as Opnion mininq. T i s used on the \web to ánaluze h e atitude bebavioY ond emational sBate at he sendey lbis applicaion 1s implehmented thvouah a cambination of NLP andctatistics bs assiabina tbe values to the text Such as posiiveneaaive, ax naBuxal and ther Yecoanize he mood ot the comtext fox example, haPpy Sad, ana4 etc 3. hat ave the ditexent challenaes in NLP Chall ergeg at NLP L Contextual woxds and phrases and homonsms Tbe same waxds_anol phxas.es can have diverse meanings accordina to the Context oT.aSentence and many aYos have fhe exact but compleely sanme pronaunLahon difexentt meaning Date Page Homcnum.smeanshe probounciotion ot two oY move woxds is same but have ditteren! meaninq. For example, their and heve viaht atnd write T6is will cxeate problem In_9ue stion ansevinq and speecb-to-text applicaions 2. Sunonums ynanums can caUse sSues like contextual undeystandiha since ale use many ditHerent wavds to expxess the identical idea Addiionally, some ot hese iwayds may Conue exactlu the same meanino, 1wbile Somemau be levels ot conaplex1ts oand. dittexet people use sunanyms to denate sliahtlu didtexent meaninqs Aithin theix pevsonal vocabulary For examplesmall itle tinu minute have meanina YOnuand_Sarcasm xonu and saYcasm pYesent pxoblems ox aachin e leaxninq models since uaualluuse aloxd s and pbxases thatthey stxicty Sycetiniti.or, mau be positive ox but rulu neaaive raean he opposite. Madels_can be Jreaind rtb. iDdicaione Hhat fyeouentlLy cextain accompalS AY.onicOrSarcastie pbyase.slke ases, lk.e yeah ueah gyiaht whateexetc and wovd t s Sila complicated ermbeddingsbu PYocess procegg Date Page 4. Ambiquity Ambiguity in NLP yetexs to senie cno phxases.that potentially hove two.ox moxe poss1ble intexpy etaiong. Ihexeislexical, syntactic and semaric ambiguity 5. EyYONS in text or speech Misspelled oY misus ed wovd.s can qenevate pxobleras dox text analusis. Auto.corxeet and aramnnaxcoYrection applicaions can handle cammon mistakes, but do not all Liraesundexstand the witex's intenion alith spoken lanquaae itis diticult fox he machine to undexstand mispronauneiations, dittexent accents, stammex.S, ete 6 Tdioms ond slana Intoymal phrases, expressionsidiam.s, and cultue-specitic linge present a number of Prablems to NLPespecially foxmodels intended ox compxehensive use Because as tornaal languaae,idi.omg may bave Do dicionaxudeinihon at all, and these expres.gions_ may even have difevent raeaninas in ditfexent geographic aYeas. Fudhexmox ecultural slanq is continouslumerphinq andncYeainq,so new wards oise evex day Date Page 7 Domain specitic lanquage Ditexent busin es.s and indusies atten Use vexu dittexent lanquaqe Hn NLP pYocessino modelneeded or healhcore would be vexy ditteent than One uS ed toprocess leaal documents 8 Low- yesaurce QnaLoaes Axiticia TntellierDce pnaehine leaxninqNLP applicahons bave been mostlu built toY Hhe most commoD widelu USedA lanauages T i s absaludely iDCYedible ot ho peige translahon Sustem bae beaormeHowevex, many lanauaaes, e.speciallu those spoken bupeople withless access -to teshnoloqy oten Ao ovevlaoked and ubder pYocessed Fox example, heye are aver 3060 lanquaaec in Aica, alane here simpl isn't ample ata on manu nese_lonquaaes