Peramalan Harga Saham Menggunakan Metode Extreme Learning Machine (ELM) Studi Kasus Saham Bank Mandiri

Muhammad Iqbal Pratama (Fakultas Ilmu Komputer, Universitas Brawijaya)
Putra Pandu Adikara (Fakultas Ilmu Komputer, Universitas Brawijaya)
Sigit Adinugroho (Fakultas Ilmu Komputer, Universitas Brawijaya)

Jurnal Pengembangan Teknologi Informasi dan Ilmu Komputer
Vol 2 No 11 (2018): November 2018

Stock investment is one of the most profitable type of investment. One of the biggest problem in stock investing is the difficultness to predict a stock price and it led to doubt whether to buy or sell a stock. Extreme Learning Machine is implemented to predict a stock price using Bank Mandiri's stock as a case study. This algorithm has some advantages such as fast training time and small error value. Extreme Learning Machine's processes involve normalizing Bank Mandiri daily stock data, generating input weight and bias weight, training the model, testing the model, denormalizing predicted value and evaluating the model using Mean Absolute Percentage Error (MAPE). The features used to predict Bank Mandiri's stock price are Open, High and Low price. The smallest MAPE value obtained from the testing phase is 1,012% using sigmoid activation function, four neurons in hidden layer and the data used is the last one year.