Repository Universitas Pakuan

Detail Karya Ilmiah Dosen

Boldson Herdianto Situmorang, Arie Qur Ania, Dian Asri Sukmawati

Judul : The Application of Prediction of Cocoa Yields with Exponential Smoothing Method
Abstrak :

— Cocoa is the one of Indonesia superior comodities
from agricultural sector, besides oil palm and rubber which
contribute the foreign exchange for the country in the amount of
US$ 1,05 billion from export of cocoa seed and cocoa product in
2012 [1]. Caused by cocoa plant is concidered important for
Indonesia economic matters, a system of prediction of cocoa
yields is needed to minimize the decrease of cocoa yields in every
year. The prediction may be used of various calculation methods,
among others by analysis method of Exponential Smoothing–a
prediction method of moving average by weighting where data is
weighted by an exponential function [2]. This research has
purpose to design and develop an application which can be used
to predict the cocoa yields by Exponential Smoothing method so
that can help the government and farmers to predict cocoa yields
per annum. The development of this application will be limited
by calculation of statistics to predict the cocoa yields base on data
in previous year. The data which will be used in this research is
the secondary data which were from Research Institute for
Industrial Plants, that are data of cocoa yields from 2008 to 2012.
The development of the application is using softwares with
XAMPP, PHP, and MySQL.  The prediction of cocoa yields in
2013 at significance level  0,5 is 17.885 Kg and it is the prediction
result with the minimum value of MSE (Mean Square Error)
4.209. After this, the results of prediction of cocoa yields per
quarterly has been a little bit change in the certain seasons. For
the validation in calculation of MAPD (Mean Absolute
Percentage Definition) the Exponential Smoothing is exactly
accurate in MAPD value at significance level 0,1 with the
confidence interval 99,427 percent, at significance level 0,3 with
the confidence interval 99,86 percent, and at significance level 0,5
with the confidence interval 99,57 percent.

Tahun : 2015 Media Publikasi : Prosiding
Kategori : Prosiding No/Vol/Tahun : - / - / 2015
ISSN/ISBN : 2461-1328
PTN/S : Universitas Pakuan Program Studi : ILMU KOMPUTER
Bibliography :

[1] F. Ahmad. “Kakao komoditi andalan Indonesia”, 2013.
[2] M. Spyros.“Metode dan Aplikasi Peramalan”, Erlangga, 2007.
[3] P. Subagyo.“Forecasting : Konsep dan Aplikasi”, BPFE Yogyakarta,
[4] S. Rahma. “Analisa trend dan forecasting”, Modul 8, 2013.
[5] D. Martisunu.“Statistik time series”, Modul 5, 2012.