Claim Missing Document
Check
Articles

Found 25 Documents
Search

Diagram Kontrol Short-Run untuk Memantau Variabilitas Proses Budi Nugroho; Desi Yuniarti; Sri Wahyuningsih
EKSPONENSIAL Vol 10 No 1 (2019)
Publisher : Program Studi Statistika FMIPA Universitas Mulawarman

Show Abstract | Download Original | Original Source | Check in Google Scholar | Full PDF (436.994 KB)

Abstract

Short run control chart is designed for production small scale and a limited amount data for monitoring different characteristic on same diagram control. The purpose of research is to know the implementation of short run control chart for monitoring mean and variability process based on characteristic data 1 inch Polyvinyl Chloride (PVC) in PT. Maspion. The objective of characteristics 1 inch pipe in this research are socket external diameter, socket internal diameter, barrel external diameter and barrel internal diameter. The Results showed that controlling process of by applying influence function, effective is to be applied to detect small movement in observation 11, 14, 14 and 9 for every control chart characteristic limited control 3σ.
Model Dinamis: Autoregressive Dan Distribusi Lag Muhajir Choir Nurahman; Sri Wahyuningsih; Desi Yuniarti
EKSPONENSIAL Vol 7 No 2 (2016)
Publisher : Program Studi Statistika FMIPA Universitas Mulawarman

Show Abstract | Download Original | Original Source | Check in Google Scholar | Full PDF (281.16 KB)

Abstract

Regression model using time series data not only use the effect of changing the independent variables on the dependent variable in the same period and for the same period of observation, but also use the period of time before. The purpose of this study was to determine the dynamic model autoregressive and distribution lag by type of infinite lag, and to know the effect of US dollar exchange rate against GDP in 1993-2013. Based on the analysis of data has that GDP and US dollar exchange rate has a rising trend pattern, and obtained by a simple regression model. But this model can not be used because of two assumptions have not been met and that there are heteroscedasticity and autocorrelation. So this model should be transformed using log, and log transformation model is obtained from a simple regression. The transportation model can be used as desiredint his model is only one assumption are not met and that there are autocorrelation. Then sub sequently estimating models and obtained Koyckas well as all assumptions are met, namely residual normal distribution, no problem heteroscedasticity and autocorrelation. Thus, the obtained dynamic distribution models also lag within finite lag types.
Regresi Nonparametrik Spline Birespon Untuk Memodelkan Persentase Penduduk Miskin dan Indeks Kedalaman Kemiskinan di Kalimantan Timur Tahun 2015 Ronald Tediwibawa; Desi Yuniarti; Memi Nor Hayati
EKSPONENSIAL Vol 10 No 1 (2019)
Publisher : Program Studi Statistika FMIPA Universitas Mulawarman

Show Abstract | Download Original | Original Source | Check in Google Scholar | Full PDF (362.731 KB)

Abstract

State of Indonesia is a developing country which has a problem, namely poverty. Poverty is a condition that is often associated with the needs, difficulties and shortcomings in the various circumstances of life. Measuring poverty in a region that can be done by looking at two indicators, namely the percentage of the poor population and the poverty depth index. This study uses 5 factors thought to affect the percentage of poor people and the depth of poverty in East Kalimantan which includes the average of the old school, the open unemployment rate, the labor force participation rate, population growth rate and the expectancy of the old school. The Data used in this study is the data year 2015, which is obtained from the Central bureau of Statistics of East Kalimantan Province. The method used is a nonparametric regression spline-response and determine the value of the optimal knots point using the Generalized Cross Validation (GCV). The best Model resulting from this research is the model with the point of optimal knot with the value of GCV of 31.14057 and R-squared of 86.47.
Pemodelan Faktor-Faktor yang Berpengaruh Terhadap Indeks Pembangunan Manusia (IPM) di Kalimantan dengan Geographically Weighted Logistic Regression (GWLR) Lili Widyastuti; Desi Yuniarti; Memi Nor Hayati
EKSPONENSIAL Vol 9 No 1 (2018)
Publisher : Program Studi Statistika FMIPA Universitas Mulawarman

Show Abstract | Download Original | Original Source | Check in Google Scholar | Full PDF (596.241 KB)

Abstract

Human Development Index (HDI) is an indicator to measure the success in building the quality of human life (community/population) and HDI can be used to see the results of the development. The average of Kalimantan HDI in 2016 has low HDI value however there is also high HDI value. Be observed from the score of those HDI, Kalimantan only has two categories those are medium and high. The statistical method used for determining the IPM model is the Geographically Weighted Logistic Regression (GWLR) method. GWLR is a local form of logistic regression in which geographic factors are considered and it is assumed that the data distributed Bernoulli are used to analyzing spatial data. This research was conducted to know the model of HDI and the factors that influence HDI in Kalimantan with GWLR using Adaptive Bisquare Kernel. The results showed that by using Adaptive Bisquare Kernel there are 56 different models for each district/city with the factors that affect the HDI in Kalimantan in 2016 vary by district/city as follows; the percentage of the poor population, the percentage of open unemployment, the percentage of the population graduated from college.
Klasifikasi Lama Masa Studi Mahasiswa Menggunakan Perbandingan Metode Algoritma C.45 dan Algoritma Classification and Regression Tree Hadi Dwi Darmawan; Desi Yuniarti; Yuki Novia Nasution
EKSPONENSIAL Vol 8 No 2 (2017)
Publisher : Program Studi Statistika FMIPA Universitas Mulawarman

Show Abstract | Download Original | Original Source | Check in Google Scholar | Full PDF (802.413 KB)

Abstract

Classification is the grouping samples based on the characteristics of the similarities and differences using target variable category. In this study, the decision tree is formed using C4.5 algorithm and Classification and regression tree (CART) algorithm to classify a student’s study period class of 2016 FMIPA UNMUL. C4.5 algorithm is a non binary classification tree where the branches of trees can be more than two on C4.5 algorithm, decision tree is established based on Entropy value. The purpose of CART algorithm is to get an accurate data as group identifier of a classification. CART can be applied in three main steps, namely the establishment of a classification tree, trimming of the classification tree, and determination of optimal classification tree. The main goal of this research is to determine factors which may effect on all predicate graduation who was graduated on 2016 using C4.5 algorithm and CART algorithm and also to know comparison accuracy of classification result by C4.5 algorithm and CART algorithm. The result showed that factors which affected the duration of all graduation using C4.5 algorithm are major (X4), region school (X5) and region origin (X3) and factors affected to the duration of all graduation using CART algorithm are major (X4) and Cumulative Achievement Index (X1). Precision classification in CART algorithm is better than C4.5 algorithm. C4.5 algorithm was able to predict with 40% accuracy while the CART algorithm has a predictive accuracy of 60%.