Open AccessImplementation of Naïve Bayes updateable with modified absolute discount smoothing on Pamekasan Regent SMS center data classification
Author(s) -
Badar Said,
Nindian Puspa Dewi
Publication year - 2019
Publication title -
journal of physics. conference series
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.21
H-Index - 85
eISSN - 1742-6596
pISSN - 1742-6588
DOI - 10.1088/1742-6596/1375/1/012029
Subject(s) - naive bayes classifier , bayes' theorem , smoothing , computer science , artificial intelligence , mathematics , statistics , machine learning , natural language processing , bayesian probability , support vector machine
Classification is a grouping process based on a predetermined class. Previous research has classified Regent Pamekasan SMS Center using Naïve Bayes and Modified Absolute Discounting (MAD) Smoothing, but the average classification accuracy is still equal to 76.83%. to improve the accuracy of classification then in this study applied Naïve Bayes Updateable by using MAD Smoothing. The classes used remain 15 classes: Education, Health, Infrastructure, Crime, Administrative Services, Sports, Government, Agriculture, Small and Medium Enterprises, Order, Weak Economy, Religion, Art and Culture, Natural Disasters, and Others. Before doing the classification process first done pre-processing such as equating characters, deletion of punctuation, restore abbreviation, translation of the local language (Madura), deletion of numbers, deletion of words that are not important in SMS, and stemming to convert into a basic word. Results Some experiments obtained an average accuracy of 78.89%, with the accuracy of one test reached 87.65%. And Naïve Bayes Updateable can increase accuracy by 2.07% with the addition of 0.47-minute classification time.