Penerapan Damerau Lavenshtein Distance dan Naïve Bayes Classifier dalam Identifikasi Spam pada Dokumen Sms dalam Bahasa Indonesia
Total View This Week0
Total View This Week0
Institusion
Institut Teknologi Perusahaan Listrik Negara
Author
Ramadhan, Reza Wahyu
Palupiningsih, Pritasari
Purwanto, Yudhy S.
Palupiningsih, Pritasari
Purwanto, Yudhy S.
Subject
Teknik Informatika
Datestamp
2022-09-26 07:26:20
Abstract :
Spam SMS is an sms document sent in the form of advertising news and other purposes in bulk. Typically like pornographic, fraudulent and gambling messages, spam displays content on a continuous basis without being asked and is often unwanted by the recipient. So an algorithm is needed to classify the SMS document. Naïve Bayes is a classification method that is quite popular in classifying a document. The disadvantage of the naïve Bayes algorithm is that there is no process to normalize the words in the SMS document to their basic word forms using the Damerau-Levenshtein Distance algorithm. The results of the algorithm also often give rise to a probability value of 0, so laplace correction / laplace smoothing is needed to avoid a probability value of 0 in the training data. Before doing text calculations, first go through preprocessing text starting from tokenizing, filtering, and stemming. Accuration value from this method is 91,5 %
Spam SMS is an sms document sent in the form of advertising news and other purposes in bulk. Typically like pornographic, fraudulent and gambling messages, spam displays content on a continuous basis without being asked and is often unwanted by the recipient. So an algorithm is needed to classify the SMS document. Naïve Bayes is a classification method that is quite popular in classifying a document. The disadvantage of the naïve Bayes algorithm is that there is no process to normalize the words in the SMS document to their basic word forms using the Damerau-Levenshtein Distance algorithm. The results of the algorithm also often give rise to a probability value of 0, so laplace correction / laplace smoothing is needed to avoid a probability value of 0 in the training data. Before doing text calculations, first go through preprocessing text starting from tokenizing, filtering, and stemming. Accuration value from this method is 91,5 %
Download
book
BibTex
Latex, Jabref
cloud_download
Original Resource
url resource Institution
