[SOLVED] Homework 2

Homework 2

DebatingwithQuanteda

Inthesecondproblemset,youwillstudytheUSpresidentialdebatesusingthe
quantedapackage.SinceRisacollaborativesoftware,youwillstartwithafew
snippetsofcodetocleanthedebatefragments,donebylastquarterQTA
students.

1) Installthequantedapackageifyouhavenotdoneso.
[Aquickintroductionintoquantedaisprovidedhere:https://cran.r-
project.org/web/packages/quanteda/vignettes/quickstart.html]

2) Usethestartingcodetoreadthepresidentialdebateobject.

3) AtextfragmentintheDEBATEScorpusobjectisaresponsetoaquestion
orarebuttalandcanbethoughtofasaself-containeddocument.Explore
towhatextenttheHeapslawappliesforTrumpvsClinton.Isitstronger
orweakerforeitherspeaker?(tip:seethecodeusedinclass).

4) Analyzetheevolutionoflexicaldiversityacrossthecandidatesbeforeand
aftertheybecametheirpartiesrespectivecandidates?[tip:theRscript
providedhasaprimaryindicatorvariableforidentifyingthefragments
thatcamefromnon-primarydebates]

5) Afterexploring4),doyouhaveahypothesiswhypatternsmaybemore
orlesspronouncedbetweenTrumpvsClinton?Howcouldyoutestthis?

6) Removestopwordsfromthecorpus.

7) Usingthetokenizefunction,constructseparatebi-gramsfortheHilary
Clinton/DonaldTrumppartsofthecorpus.Tabulatethetenmost
frequentbigramsbyspeaker.Aretheseinformative?Whyorwhynot?

8) Usingthecollocationfunctioninquanteda(whichtakesatokenizeobject
asargument),constructcollocationsbasedonChi2testforeachspeaker.
OrderbyChi2teststatistic.Whatdoyounoticeorwhatisstrange?Can
youprovideaformalreasoningrelatingtotheChi2teststatisticformula?

9) Usingthecodeprovidedinthelectures,identifycollocationsthatare
distincttoTrumpvsClinton.Basedonyourperceptionofeachcandidate,
dotheresultsofthisanalysismakesense?Provideabriefanswerwhere
yousummarizethemostimportantresultsandyourexplanations.

SendyourPDFversionoftheMarkdowndocumentto

[email protected]

Pleaseusethefollowingnamingconvention:HW1_Surname_FirstName.pdf
Deadline:byFriday28thApril,2017.