Open AccessPENERAPAN DATA MINING UNTUK PRAKIRAAN CUACA DI KOTA MALANG MENGGUNAKAN ALGORITMA ITERATIVE DICHOTOMISER TREE (ID3)
Author(s) -
Broto Poernomo T.P,
Rina Dewi Indah Sari
Publication year - 2017
Publication title -
joutica
Language(s) - English
Resource type - Journals
eISSN - 2621-511X
pISSN - 2503-071X
DOI - 10.30736/jti.v2i2.68
Subject(s) - outlier , id3 , computer science , data mining , process (computing) , data set , set (abstract data type) , tree (set theory) , decision tree , artificial intelligence , decision tree learning , mathematics , programming language , operating system , mathematical analysis
In the preparation of weather forecasts information there are several obstacles, involving many sources of data and weather forecasts relying on the ability of the forerunner, so that the interpretations produced may differ between forecasters because of their own experience. Differences in interpretation can confuse users and potentially reduce the quality of information submitted. Based on this problem the author intends to study the forecast model with data mining using ID3 algorithm to obtain the appropriate model so as to facilitate the process of weather analysis and forecast. In building the application, data obtained from BMKG website is addressed dataonline.bmkg.go.id and the data taken is datacuaca on January 1, 2012 through November 30, 2015 for a total of 2414 data and 9 attributes. After the selection of attributes (only weather related attributes) and removing damaged data (incomplete data and outliers), the data is reduced to 1790 and attributes reduced to 6 pieces. In the testing process performed with 179 data (10% of the data set) random, it is known that there are 112 databases corresponding to the actual weather data. So it can be concluded that the accuracy of 73.74%.