ENGINEERING JOURNAL : PENGEMBANGAN ANT ALGORITHM DENGAN MENGGUNAKAN KONSEP GLOBAL DESIRABILITY DAN GLOBAL FREQUENCY PADA DATA MINING UNTUK PENGKLASIFIKASI DATA

Jurnal Rekayasa Vol 1 No 2 Artikel 5

Saiful Bukhori
(Staf Pengajar Jurusan Elektro, Program Studi Teknik, Universitas Jember)

ABSTRACT

Application of ant algorithm for solving the problem with artificial intelligence has grew at full speed. This problem was related to the behavior of ant that was aimed at strove for sustain the ant colony life. Base on the literature survey, the research that was apply the ant algorithm in data mining scope for data classification has not done. In this research, researcher was design and analysis application of ant algorithm in data mining for data classification. The algorithm, that was used, was modification and improvement for the algorithm that has developed previously. Ant algorithm that was designed was influenced four main parameters, ant desirability, ant frequency, heuristic information () and pheromone concentration () The result of the design and analysis algorithm was ant desirability and ant frequency parameters could reduce the time complexity from O(mn2) to O(mn). The Software that was designed and implemented in operating system Windows has experimented with data from many source. The experiment was done with many configuration of parameters. The result of the experiment was the true classification in 93,48% - 97,00% range, the false classification in 3,00% - 6,52% range and unclassified 0%.

Key Word: data mining, data classification, ant algorithm, ant desirability, ant frequency

DOWNLOAD NASKAH LENGKAP:

ZIDDU:

http://www.ziddu.com/download/7522913/vol_1_no_2_artikel_5_Saiful_Bukhori.pdf.html

SEMOGA BERMANFAAT