Create a tokenizer

Write a function tokenize that takes a string and returns a list of tokens.

[ ] deftokenize(doc): return

Calculate token scores

Calculate scores for every token in the corpus, using the method discussed in class. Store these scores in a dictionary called token_scores.

[ ] token_scores={}

Create a score message function

Write a function score_message that takes an SMS message doc and returns a SPAM score, using the method discussed in class.

[ ] defscore_message(doc):

return

1. What tokens are most predictive of a message being SPAM? (coding)
2. What tokens are most predictive of a message being HAM? (coding)
3. How many documents are misclassified by the model?(coding)

import urllib.request, json 

sms_corpus = []

with urllib.request.urlopen("https://storage.googleapis.com/wd13/SMSSpamCollection.txt") as url:

  for line in url.readlines():

    sms_corpus.append(line.decode().split('t'))

 

# print the text and label of document 10

docid = 16

print(sms_corpus[docid])

 

# print the label of document 10

docid = 16

print(sms_corpus[docid][0])

 

# print the text of document 11

docid = 16

print(sms_corpus[docid][1])

1. Can you get better results by improving your tokenizer?

Transcribed Image Text:

- SMS SPAM Collection The SMS SPAM Collection is a corpus of real text messages (SMS messages) that have been classified as either SPAM or HAM (i.e. not SPAM). The corpus contains 5,574 documents, 747 of which are SPAM and 4,827 of which are HAM. You can find the readme for the corpus here. The following code downloads a copy of the SMS SPAM Corpus and saves it in a variable sms_corpus. import urllib.request, json sms_corpus = [] with urllib.request.urlopen ("https://storage.googleapis.com/wd13/SMSSpamCollection.txt") as url: for line in url.readlines(): sms_corpus.append(line.decode().split('\t')) sms_corpus is a list. Each element of the list is another list which stores a document and its label. [ ] # print the text and label of document 10 docid= 16 print(sms_corpus[docid]) ["ham", "Oh k...i'm watching here:) "] [] # print the label of document 10 docid= 16 print(sms_corpus[docid] [0]) ham [] # print the text of document 11 docid= 16 print (sms_corpus[docid][1]) Oh k...i'm watching here:) - SMS SPAM Collection The SMS SPAM Collection is a corpus of real text messages (SMS messages) that have been classified as either SPAM or HAM (i.e. not SPAM). The corpus contains 5,574 documents, 747 of which are SPAM and 4,827 of which are HAM. You can find the readme for the corpus here. The following code downloads a copy of the SMS SPAM Corpus and saves it in a variable sms_corpus. import urllib.request, json sms_corpus = [] with urllib.request.urlopen ("https://storage.googleapis.com/wd13/SMSSpamCollection.txt") as url: for line in url.readlines(): sms_corpus.append(line.decode().split('\t')) sms_corpus is a list. Each element of the list is another list which stores a document and its label. [ ] # print the text and label of document 10 docid= 16 print(sms_corpus[docid]) ["ham", "Oh k...i'm watching here:) "] [] # print the label of document 10 docid= 16 print(sms_corpus[docid] [0]) ham [] # print the text of document 11 docid= 16 print (sms_corpus[docid][1]) Oh k...i'm watching here:)