Realized Volatility Forecasting for AI-Mining Sensor Data Using the Multi Layer Perceptron Method

Obed Christian Dimitrio (Universitas Kristen Satya Wacana)
Didit Budi Nugroho (Universitas Kristen Satya Wacana)
Hanna Arini Parhusip (Universitas Kristen Satya Wacana)
Atyanta Nika Rukmasari (Universitas Kristen Satya Wacana)

Journal of Science and Science Education
Vol 5 No 2 (2021): JoSSE Vol. 5 No. 2 (November 2021)

This study aims to predict the Realized Volatility (RV) value from AI-Mining sensor data for the period 23 May to 6 June 2022 by using the Multi Layer Preceptron (MLP) method. MLP is the simplest method of artificial neural network. Based on the results obtained after doing MLP with the Python language on Google Colab, the predicted RV value for each data shows a movement in value that is almost similar to the original RV value. The Root Mean Squares Error (RMSE) value for each data prediction is relatively small, which indicates that the MLP method provide accurate prediction on the use of the AI-Mining sensor data to forecast RV.