|
@@ -98,14 +98,14 @@ for (sm in spam_messages) {
|
|
|
words <- str_split(sm, " ")[[1]]
|
|
words <- str_split(sm, " ")[[1]]
|
|
|
spam_vocab <- c(spam_vocab, words)
|
|
spam_vocab <- c(spam_vocab, words)
|
|
|
}
|
|
}
|
|
|
-spam_vocab <- spam_vocab
|
|
|
|
|
|
|
+spam_vocab
|
|
|
|
|
|
|
|
ham_vocab <- NULL
|
|
ham_vocab <- NULL
|
|
|
for (hm in ham_messages) {
|
|
for (hm in ham_messages) {
|
|
|
words <- str_split(hm, " ")[[1]]
|
|
words <- str_split(hm, " ")[[1]]
|
|
|
ham_vocab <- c(ham_vocab, words)
|
|
ham_vocab <- c(ham_vocab, words)
|
|
|
}
|
|
}
|
|
|
-ham_vocab <- ham_vocab
|
|
|
|
|
|
|
+ham_vocab
|
|
|
|
|
|
|
|
# Calculate some important parameters from the vocab
|
|
# Calculate some important parameters from the vocab
|
|
|
n_spam <- spam_vocab %>% length()
|
|
n_spam <- spam_vocab %>% length()
|
Casey Bates