Prediksi Pertumbuhan Penduduk di Kota Malang menggunakan Metode Extreme Learning Machine (ELM)

Inas Hakimah Kurniasih (Fakultas Ilmu Komputer, Universitas Brawijaya)
Muhammad Tanzil Furqon (Fakultas Ilmu Komputer, Universitas Brawijaya)
Sigit Adinugroho (Fakultas Ilmu Komputer, Universitas Brawijaya)

Jurnal Pengembangan Teknologi Informasi dan Ilmu Komputer
Vol 4 No 2 (2020): Februari 2020

Dinas Kependudukan dan Pencatatan Sipil (Dispendukcapil) in Malang City is tasked with provide public services in terms of civil registration such as making electronic Kartu Tanda Penduduk (e-KTP) and birth certificates. Dispendukcapil prepared this by planning for the target needs and predicting the population in the next 5 years, but the error value is unknown. This research helps predict with a small error value using Extreme Learning Machine (ELM) method and calculates the error value using Mean Absolute Percentage Error (MAPE). Based on the results of testing implementation and analysis, using data from 2009 to 2019 obtained MAPE error value of 0.498% and runtime 1.166 seconds with the use of 3 input neurons, 5 hidden neurons, binary sigmoid, as well as 50 training data and 66 testing data. Then, in implementation of testing using data from 2012 to 2019 obtained MAPE error value of 0.117% and runtime 1.227 seconds with the use of 3 input neurons, 6 hidden neurons, binary sigmoid, as well as 70 training data and 4 testing data.