Page 85 - Conference Book The 3rd IConVET "Future of TVET Graduates: Developing Talent for Industry 4.0 and The New Normal"

The 3 International Conference
rd
on Vocational Education and Technology (IConVET)
Universitas Pendidikan Ganesha - Bali | 7 November, 2020
th
FORECASTING FOREIGN EXCHANGE RATE USING
A COMBINATION OF LINEAR REGRESSION AND
FLOWER POLLINATION ALGORITHM

I B N Pascima, and I M Putrama

Faculty of Engineering and Vocational, Universitas Pendidikan Ganesha,
Indonesia

gus.pascima@undiksha.ac.id, made.putrama@undiksha.ac.id

Abstract. Several currencies exist in the world. Each currency will have
value. The currency exchange will go through a conversion process to adjust
the amount. Each currency value can fluctuate based on the conditions of the
currency area. The fluctuating changes in value provide profit opportunities.
Maximizing profits can make forecasts so that the right decisions are made.
One of the forecasts can use regression. Regression is capable of forecasting
based on historical data. The regression in this study will be optimized using
the Flower Pollination Algorithm (FPA). The use of the Flower Pollination
Algorithm (FPA) aims to obtain appropriate parameters for regression to
reduce forecast errors. The data in this study were obtained utilizing
extraction from the Meta trader application. This data will be the basis for
the system learning stage and the testing phase. Obtaining a good
hyperparameter can make the forecasting system closer to the actual value.
Good system accuracy can be a trader's supporting data in making
transactions. Forecasting in this study uses the parameter 5 window sizes, 20
population sizes, and 0.7 probability switch. This experiment resulted in
MSE 0.0331 and RMSE 0.1756. This forecasting has sufficient results to
support a trader's decision. Further research is needed to improve accuracy
and determine the direction of the forecast to improve this research.

| IConVET 2020 | 85

https://conference.undiksha.ac.id/iconvet/ | 85 |