<![CDATA[Abstract- The information via the internet spreads rapidly. Not only does it have a positive impact, but also a negative impact, one of which is the ease of finding information that contains negative things such as pornography. That problem can be overcome by research on machine learning-based classification of pornographic texts using the Naive Bayes method. Naïve Bayes is a probabilistic classifier that calculates a set of probabilities based on the number of frequencies and combinations of each value from a given dataset. Through Naïve Bayes, it is expected that there is a possibility to classify the text of story titles containing pornography downloaded via a web page. The type of data used is pornographic story titles downloaded via porn sites. There are 100 story title datasets employed. The classification stages using the Naive Bayes Classifier Algorithm include Preprocessing, TF-IDF weighting, Naive Bayes classification, and testing using Classification Accuracy, Precision, and Recall. The test results using the Data Splitting method are examined into five comparisons with TF-IDF values with each level of Accuracy, Precision, and Recall up to 76.44%, 82.43%, and 73.43%. The characterization without using TF-IDF, respectively, obtained Accuracy, Precision, and Recall values up to 97%, 96%, and 93%. The result is that the Naïve Bayes technique without TF-IDF creates a more favorable execution than TF-IDF. Keywords: pornographic story titles, classification, Naive Bayes]]>
Copyrights © 2025
