Write a function that preprocesses the natural language data and returns the stems of the tokenized tweet. $(15\%)$ a$.$ Remove the Twitter handles $($i$.$e$.$ @user$\_$id$)$ in the column tweet b$.$ Remove punctuations, including $",.$;:$?!$#$"$ c$.$ Remove numbers, i$.$e$."0-9"$ d$.$ Remove special characters and non$-$English characters e$.$ Remove words with length $3$ f$.$ Tokenize the comments g$.$ Apply stemming on the tokens and return the stems of the tokenized tweet. Code:

$```$

# Problem $3$

import re

import nltk

from nltk$.$stem.porter import $*$

import string

def problem$\_3($df$)$:

# write your logic here

# tweet data is stored in df$[\text{'}$tweet$\text{'}]$

tokenized$\_$tweet $=[]$

# remove Twitter handles

# remove punctuations $,.$;:$?!$#

# remove numbers

# remove special characters

# remove short words, length is regarded as short word

# tokenization and stemming

return tokenized$\_$tweet

$```$

Execution: $\backslash (\backslash$square $\backslash )$

$```$

$>$ df $=$ pd$.$read$\_$csv$("$tweet$\_$data.csv$")$

$>$ tokenized$\_$tweet $=$ problem$\_3($df$)$

$>$ print$($tokenized$\_$tweet$)$

$0[$when$,$ father, dysfunct, selfish, drag, kid, i$...$

$1[$thank$,$ lyft$,$ credit, can't, caus, they, don't...

$[$bihday$,$ your, majesti$]$

$[$model$,$ love, take, with, time$]$

$[$factsguid$,$ societi, motiv$]$

$*.$

$[$that$,$ youuu$]$

$[$nina$,$ turner, airwav, tri, wrap, herself, man...

$[$listen$,$ song, monday, morn, work$]$

$[$sikh$,$ templ, vandalis, calgari, condemn$]$

$[$thank$,$ follow$]$

tidy$\_$tweet, Length: $31962,$ dtype: object

Name: ting$($tokenized$\_$tweet.shape$)$

$(31962,)$

$```$