0 ratings0% found this document useful (0 votes) 62 views5 pagesNLP Module 1 Introduction
Natural Language Processing Notes University of Mumbai computer engineering
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content,
claim it here.
Available Formats
Download as PDF or read online on Scribd
Medi: dtho Netweal Language Rrocesing
NP 2 de fochnolo rod by machines +4 understand, analyse g
™anfcul ate human languages
* Phases 4 NLP Dependency
J Pos Tasos
nding” POS Tagatng Word sence
NZ V Discrlseute
aa | Zemantec dpe
ut Sentence\—»|Lateat | —s|Ayatax 7
— ee | Ants fags PSeerle
Rule,
Toherijatton Stop, -RMemufing 2
lohedigal Sth, Lenmatidet
Remeval
aconterdug(
Flo
Anoluss |
0 hereal Anal <2 CMoxphelogical)
Words ave ap seperated ato tohens B these tokens axe reduced,
to dhete root dorms
Ke
=D Toberigalton — sptitfing by spaces or sentences
9 Bon word Removal —> Commonly used *uselace" worls Ite Cand, ef, tho)
> reduces nefse
9 Btorating —» Reduces words to voot by directly vemafing suffices fied
* It seco check Y base words rates any Sense or not
changing, changed, change —> chang
d) temmatgalion + Lome Reduction but $i actually Crede @ Acwar
Worn
changing changed, changes change© Banter nays
> Analyse de Grammattcal ahuctue af sentences fo devtve eve
Oord's vole to g Seatence
d
* Dependency Pain
& bostcally de velakonchip of a Word with another as
@ subfect
olfect
suly moditier
ok modifier
S
Deomed Universities charge huge decg CO ere oe
x 4 es Weer adh"
LT = aces stipe
Mo:
% Assign
IS Grammalfea) tags Coun, verb) ete fo du word baced
Gn qe sentence
® Semontfe Analyste
> Check 4 du seateace mabes Gay Sease or not
> Understand ate mearting of dle geatence
9 Weed. Lease Drscmbigatin
& ddenkegy cowed monsting q oord $# mullfple we oaste
"Sam went to Ange’ “eb ate de -oppleg
company Feat
0 Nomad ay Race
% ddortify g classy nome 4 org, lace or (ecpie
“4 lfue @a Now Bombay *ex4uRey AP canted of a word by tee gato consfderatton
fe whole text
&
Aettences Centeay Peeneune Need! dhis
uy
OFogmate Arar
~ Bastaaly understand fuman "Sarcasm" go macttne chould be
able {, uaderctand Ympled measing beyond. tS Meeal meading
Exceptes :
“Hey tohy ave you Se early, ou should. tobe your ocoa Rove“
“Can you pass da salt 2"
sto Meceingy
\
YES/NO Questor but ia veality
its & xeqpest fo och be salt
| Impled roca
“Please pose me de salt"
K[Ambiguties § its Tynes
"Ambe guilty beetoyl means a sentence /phrase con be talerpreted
0 eulkple ways
© Lercal Prob uty CEh word ke do mearing)
> Gandk Cs@x or rtuer bab)
> OF Cyesterday oy tommorow
= Ai € FT Checblace ov desert land)
7© Ayatacke
~ Sentence can i
—
(lased #n multiple ways
She so de mon uth o telescopa"
fe man She used dt
hod a telescope felowoge be Me See
dhe man
> 88 St on a or
f \
oatched F was deenting
hier chafing tea 2 { Saahim
tea
© Seoemantie
> meastng od de pease 4s aerbTofous
— “fhe chrchen 7 “er
“ Cooked clfcber Lavo. chicken %
we can eat geting be oat sth
© Relorendtal Ambtgaity
> the pronaias potat to whch acu .
“John told Mate that he % gs fo de stove
mee
could GE Juha ov mosh
© Pragmatic
—> "gts cold Pn here, esn’t #2
/ \
feliag cheat dadrecly ask
q <
tempera! ve fe Gncvese tem* Challenges #9 NUP
Oo Ambaguttes
® Understandia 14 Context —> Sarcam, Geom, Humour
© Language Voxtalt thy —> Slangs, comlacd languages
Citinglish,)
® Low Resource Languges —> Tulu, Hoobly less data te
ges 1
heen or
© Complex Rentences wht Long Term Dependonfes
© Lithecal Coneecleralfon g Pivacy
¥ Aaplecatfors of NLP
© Search Engines — Goce
® Frarglatfon Ages — Google Translate
® Speech RocogritRon— Si £ Goegle Acsstaut
® Chabbots — epr
© Lonltment Anolss —> veins
—@ Onttcak Character Recoantelfon — used to work on Scanned
documents
@ Autocomplete & Auto comrect
® Text Surrmast elfen & Extraction