Article Per Year (5 Year)
p-Index From 2019 - 2024
0.23
P-Index
This Author published in this journals
All Journal Jurnal Pengembangan Teknologi Informasi dan Ilmu Komputer
Isti Marlisa Fitriani
Fakultas Ilmu Komputer, Universitas Brawijaya
Computer Science & IT Control & Systems Engineering Education Electrical & Electronics Engineering Engineering

Published : 1 Documents Claim Missing Document
Claim Missing Document
Check
Articles

Found 1 Documents
Search

 1 
Klasifikasi Senyawa Kimia dengan Notasi Simplified Molecular Input Line Entry System (SMILES) menggunakan Metode Extreme Learning Machine (ELM) Isti Marlisa Fitriani; Dian Eka Ratnawati; Syaiful Anam
Jurnal Pengembangan Teknologi Informasi dan Ilmu Komputer Vol 3 No 5 (2019): Mei 2019
Publisher : Fakultas Ilmu Komputer (FILKOM), Universitas Brawijaya

Show Abstract | Download Original | Original Source | Check in Google Scholar | Full PDF (318.401 KB)

Abstract

Indonesia has a huge natural's potential by the existence of various plants and animals discovery. This issue brings a good for Indonesian people through taking advantage of nature, especially in pharmacology. In pharmacology, active compounds can be used to prevent and cure disease. Therefore, a research is conducted in informatics's field by making an active compounds' classification system to determine its pharmacological benefits. SMILES is a chemical compound notation used in this research. SMILES's features which are used as many as 15, namely B, C, N, O, P, S, F, Cl, Br, I, OH, @, =, #, and (). ELM is an ANN method that can do a generalization better than conventional methods in a limited time. A number of hidden neurons test which were conducted using k-fold cross validation method in 2 classes produced the best accuracy, 85,03%, in Metabolism and Inflammation class scenario with a total of 5, 10, and 15 hidden neurons. A number of hidden neurons' test use k-fold cross validation method which were conducted in 3 classes produced the best accuracy, 55,06%, in Metabolism, Inflammation, and Cancer class scenario with a total of 300 hidden neurons. The best accuracy was obtained as many as 55,06% by testing 15 features with 300 hidden neurons, while in 11 features's test with 400 hidden neurons was found a number of 49,18% as the best accuracy.
Co-Authors Anam, Syaiful Dian Eka Ratnawati