Implementing Hidden Markov Model to Predict Foreign Exchange Movement

Autor:	Weny Mistarika Rahmawati, Tri Swasono Himawan, Tutuk Indriyani
Rok vydání:	2018
Předmět:	Absolute deviation Currency Econometrics Economics General Medicine Foreign exchange Forward algorithm Hidden Markov model Foreign exchange market Profit (economics) Error score
Zdroj:	INTEGER: Journal of Information Technology. 3
ISSN:	2579-566X 2477-5274
DOI:	10.31284/j.integer.2018.v3i1.140
Popis:	Investment refers to personal bussiness. So many people have got profit from investment both real and non real sectors. Foreign Exchange (FOREX) is the example of non real sector. The currency fluctuation of FOREX usually occurs and this causes many investors fooled by the pattern of currency fluctuation. Finally, they get lost and even lost capital. Hidden Markov Model was implemented in this research to predict the movement of FOREX of 8 currencies. The data were trained by Baum-Welch algorithm and predicted by Forward algorithm. The trial obtained the average MAPE (Mean Absolute Precentage Error) of 8 currencies which was relatively small (0.0038082% belongs to high and 0.0040706% belongs to low), less than 1%. The currency of USD/IDR has the smallest error score among the other tested currencies. Its average MAPE was 0.0032624% and the average deviation was 42. Thus, this system is well proven to predict the movement of currency.
Databáze:	OpenAIRE
Externí odkaz:	https://explore.openaire.eu/search/publication?articleId=doi_________::37425af3d8de85600fcf3753bb384b9f https://doi.org/10.31284/j.integer.2018.v3i1.140 Zobrazit plný text záznamu