Kajian Panjang Data Historis yang Representatif pada Model Stokastik

*Setiarso Gunawan -  Magister Teknik Sipil Program Pasca Sarjana Universitas Diponegoro
Sri Eko Wahyuni -  Jurusan Teknik Sipil Fakultas Teknik Sipil Universitas Diponegoro, Indonesia
Suharyanto Suharyanto -  Jurusan Teknik Sipil Fakultas Teknik Sipil Universitas Diponegoro, Indonesia
Published: .
Open Access
Citation Format:
Article Info
Section: Articles
Language: EN
Full Text:
Statistics: 142 593
Abstract

Stochastic models are models to generate new data series based on historical data and have similar statistical parameter with statistic historical data. Methods of forecasting are developed base on statistic and mathematic science. The historical data are observed data or sample data. The limited data is become main constrain for extrapolation of data. The mean error of generated data should be lower than 5%, its mean data of generated have the validation rate on 95 %. Three samples location for study are Catchment of Bengawan Solo in Bojonegoro, Catchment of Serang in Kedungombo - Grobogan and Catchment of Citarum in Cirata - Bandung. The synthetic data and then is used to calculate the statistic parameter. Error of generated data is measured with relative error. The relative error is result of divided and subtract statistic parameter of generated data and the statistic parameter of historical data longest and statistic parameter of generated data. The result of data length analysis is relative error and historical length of the data. The analyzed result indicate that historical data are studied have representative historical data about 30 years length of data.

Keywords: stochastic, historical data, synthetics data, representative data length and relative error

Permalink: http://www.ejournal.undip.ac.id/index.php/mkts/article/view/2093

[How to cite: Gunawan, S., Wahyuni, S.E. dan Suharyanto, 2006, Kajian Panjang Data Historis yang Representatif pada Model Stokastik, Jurnal Media Komunikasi Teknik Sipil, Volume 14, Nomor 2, pp. 129-141]

Article Metrics: