Tatik Widiharih
Jurusan Statistika FSM Undip

Published : 51 Documents
Articles

PENDEKATAN REGRESI POLINOMIAL ORTHOGONAL PADA RANCANGAN DUA FAKTOR (DENGAN APLIKASI SAS DAN MINITAB) widiharih, Tatik
MATEMATIKA Vol 4, No 1 (2001): JURNAL MATEMATIKA
Publisher : MATEMATIKA

Show Abstract | Download Original | Original Source | Check in Google Scholar | Full PDF (56.697 KB)

Abstract

            Pendekatan regresi polinomial orthogonal dapat dilakukan pada rancangan dengan faktor kuantitatif dan jarak antar taraf faktor sama. Pendekatan ini dilakukan bila peneliti ingin menentukan taraf faktor dari masing-masing faktor yang mengoptimalkan respon yang diamati. Penentuan derajat polinomial berdasarkan kontras-kontras orthogonal yang nyata (significant) dari masing-masing faktor, kemudian dapat ditemukan bentuk regresi pendekatannya. Persamaan yang diperoleh merupakan fungsi matematika dengan dua peubah. Dengan menggunakan hitung differensial dapat ditentukan titik ekstrem dari fungsi tersebut. Bila dikembalikan kebentuk rancangan berarti dapat ditentukan taraf faktor dari masing-masing faktor yang mengoptimalkan respon yang diamati.
ESTIMASI DATA HILANG PADA RANCANGAN ACAK KELOMPOK LENGKAP widiharih, Tatik
MATEMATIKA Vol 10, No 2 (2007): JURNAL MATEMATIKA
Publisher : MATEMATIKA

Show Abstract | Download Original | Original Source | Check in Google Scholar | Full PDF (73.893 KB)

Abstract

Randomized complete block design is a design to reduce the residual error in an experiment by removing variability due to a known and controllable nuisance variable. Missing observations introduce a new problem into the analysis since treatments are no longer orthogonal to blocks, that is, every treatment does not occur in every block, There are two general approaches to the missing values problem. The first is an exact analysis, the second is an approaximate analysis in which the missing observations are estimated and usual analysis of variance is performed just as if the estimated observations were real data, with the error degrees of freedom reduced by the number of missing observations. In this paper was discussed the second approach with completely analysis. Bigger?s method is a simple method for estimating missing observations by using matrix approximation.  
METODE TAGUCHI UNTUK OPTIMALISASI PRODUK PADA RANCANGAN FAKTORIAL Wuryandari, Triastuti; Widiharih, Tatik; Anggraini, Sayekti Dewi
MEDIA STATISTIKA Vol 2, No 2 (2009): Media Statistika
Publisher : Department of Statistics, Faculty of Science and Mathematics, Universitas Diponegoro

Show Abstract | Download Original | Original Source | Check in Google Scholar | Full PDF (430.172 KB) | DOI: 10.14710/medstat.2.2.81-92

Abstract

Taguchi methods represent the effort quality improvement which known as off-line quality control  method because the method design quality into every appropriate process and product. Taguchi methods is represent quality repair with attempt ?new? methods, its meaning do dissimilar approach giving same belief storey by SPC (Statistical Proces Control), very effective in quality improvement as well as lessening expense of same. Fractional factorial design represent base from Taguchi method by fraction from factorial design. Fractional factorial with  4 factors and defining relations p = 2 is or 81 run become or 9 blocks with each blocks there are 9 run just eligible one block. The block name that is Orthogonal Array which lessen time and attemp fare. Orthogonal Array used to device of factorial attemp 3 level by 4 factors that is Orthogonal Array L9. Optimalitation product of factorial design  can be determinate with tables of anova, table of response and tables of Signal to Noise Ratio.   Keywords: Taguchi Methods, Signal to Noise Ratio, Orthogonal Array
PENDEKATAN SISTEM PERSAMAAN SIMULTAN DALAM PEMODELAN PRODUK DOMESTIK REGIONAL BRUTO (PDRB) PROPINSI JAWA TENGAH TAHUN 2000-2010 Satrio, Rizky Oky Ari; Widiharih, Tatik; Hoyyi, Abdul
Jurnal Gaussian Vol 1, No 1 (2012): Jurnal Gaussian
Publisher : Departemen Statistika FSM Undip

Show Abstract | Download Original | Original Source | Check in Google Scholar

Abstract

Gross Domestic Product (GDP) is general indicator used to identify the economical development in a region. The condition of economy in Central Java Province is categorized as stable condition since it has GDP value developed rapidly year by year. Refer to model used by Bappenas,the simultaneous equation model between GDP is influenced by number of employee and government spending.Identification of the model in this study using the ordercondition of indetification on the basis of the result of the overidentified taken the GDP of agriculture, mining, electricity, gas and water sector and trade. Therefore, the parameter evaluation used is 2SLS method (Two Stage Least Square). After fulfilled  assumption of independent, identical and normal distribution, the most valued toward model of GDP in Central Java Province is GDP sector of agriculture.
MODEL LOGIT KUMULATIF UNTUK RESPON ORDINAL Raharjanti, Robiah Peni Raharjanti; Widiharih, Tatik
MATEMATIKA Vol 8, No 3 (2005): JURNAL MATEMATIKA
Publisher : MATEMATIKA

Show Abstract | Download Original | Original Source | Check in Google Scholar | Full PDF (63.836 KB)

Abstract

Logit cumulative model is used to discribe the relationship between a response variable and one or more explanatory variables which response variable is of ordinal scale. To estimate the parameters, use maximum likelihood method with Newton Raphson iteration. Testing for the significance of the coefficients is done to fit the model. Test for overall significance of the variables in the model is performed by likelihood ratio test and test on individual coefficient is done using Wald?s test
PENANGANAN MULTIKOLINEARITAS (KEKOLINEARAN GANDA) DENGAN ANALISIS REGRESI KOMPONEN UTAMA Widiharih, Tatik
MATEMATIKA Vol 4, No 2 (2001): JURNAL MATEMATIKA
Publisher : MATEMATIKA

Show Abstract | Download Original | Original Source | Check in Google Scholar | Full PDF (51.079 KB)

Abstract

Multikolinearitas yang tinggi diantara peubah-peubah bebas, mengakibatkan pendugaan dengan metode kuadrat terkecil tidak dapat diandalkan, ditandai dengan koefisien regresi tidak nyata dan adanya multikolinieritas. Pendeteksian multikolinearitas dapat dilakukan secara informal salah satunya dengan koefisien korelasi lenear antar peubah bebas maupun dengan cara formal dengan faktor inflasi ragam. Analisis regresi komponen utama digunakan untuk menghilangkan multikolinieritas dan semua peubah bebas masuk dalam model, analisis regresi ini merupakan teknik analisis regresi yang dikombinasikan dengan teknik analisis komponen utama . Analisis komponen utama bertujuan menyederhanakan peubah yang diamati dengan mereduksi dimensinya , hal ini dilakukan dengan jalan menghilangkan korelasi diantara peubah-peubah melalui transformasi . Teknik analisis komponen utama dijadikan sebagai tahap analisis antara untuk memperoleh hasil akhir dalam analisis regresi.
RANCANGAN D-OPTIMAL UNTUK MODEL EKSPONENSIAL GENERAL Widiharih, Tatik; Haryatmi, Sri; Gunardi, Gunardi
MEDIA STATISTIKA Vol 7, No 2 (2014): Media Statistika
Publisher : Department of Statistics, Faculty of Science and Mathematics, Universitas Diponegoro

Show Abstract | Download Original | Original Source | Check in Google Scholar | Full PDF (457.702 KB) | DOI: 10.14710/medstat.7.2.71-76

Abstract

Exponential model is widely used in biology, chemistry, pharmacokinetics and microbiology. D-optimal criteria is criteria with the purpuse to minimize the variance of  the estimator of parameters in the model. In this paper will discuss the D-optimal design for the generalized exponential model with  homoscedastics  errore assumtion. We used minimally supported design with the proportion of  each design point is uniform. The optimization is used  modified Newton, and the results obtained that the  design points are  interior points of the design region. Keywords: D-Optimal, Generalized Exponential, Minimally Supported Design, Support Point, Homoscedastics
RANCANGAN D-OPTIMAL MODEL MICHAELIS MENTEN DAN EMAX DENGAN MATLAB Widiharih, Tatik; Haryatmi, Sri; Gunardi, Gunardi; Wilandari, Yuciana
MEDIA STATISTIKA Vol 8, No 2 (2015): Media Statistika
Publisher : Department of Statistics, Faculty of Science and Mathematics, Universitas Diponegoro

Show Abstract | Download Original | Original Source | Check in Google Scholar | Full PDF (228.531 KB) | DOI: 10.14710/medstat.8.2.69-80

Abstract

Michaelis Menten and Emax models  are  widely used in chemistry, pharmacokinetics and pharmacodynamics areas. D-optimal criteria is criteria with the purpuse to minimize the variance of the estimator of parameters in the model. In this paper will discuss the D-optimal design for Michaelis Menten and Emax models with  homoscedastics  error assumtion.  Determination of D-optimal designs based on Generalied Equivalence Theorem Kiefer-Wolvowitz. We used minimally supported design with the proportion of  each design point is uniform, lower bound of design region is design point and the others are interior points.Keywords: D-optimal, Michaelis Menten, Emax, Minimally Supported Design, Homoscedastics
INFERENSI DATA UJI HIDUP TERSENSOR TIPE II BERDISTRIBUSI RAYLEIGH Widiharih, Tatik; Mardjiyati, Wiwin
MEDIA STATISTIKA Vol 1, No 2 (2008): Media Statistika
Publisher : Department of Statistics, Faculty of Science and Mathematics, Universitas Diponegoro

Show Abstract | Download Original | Original Source | Check in Google Scholar | Full PDF (125.239 KB) | DOI: 10.14710/medstat.1.2.69-74

Abstract

Abstrak Analisis data uji tahan hidup merupakan salah satu teknik analisis statistika yang banyak digunakan di bidang industri dan kesehatan. Data waktu hidup dapat berupa data lengkap atau data tersensor, dan merupakan variabel random nonnegatif. Estimator titik untuk parameter q digunakan MLE, kemudian MLE tersebut digunakan untuk uji kecocokan distribusi Rayleig dengan metode Anderson Darling. Estimator titik uji tahan hidup meliputi rata-rata waktu kegagalan (mean time to failure / MTTF), fungsi kegagalan h(t), dan fungsi ketahanan S(t). Estimasi interval dilakukan dengan metode besaran pivot.   Kata kunci: data tersensor, MLE, rata-rata waktu kegagalan, fungsi                      kegagalan, fungsi ketahanan
MODEL PENILAIAN KREDIT MENGGUNAKAN ANALISIS DISKRIMINAN DENGAN VARIABEL BEBAS CAMPURAN BINER DAN KONTINU Mukid, Moch. Abdul; Widiharih, Tatik
MEDIA STATISTIKA Vol 9, No 2 (2016): Media Statistika
Publisher : Department of Statistics, Faculty of Science and Mathematics, Universitas Diponegoro

Show Abstract | Download Original | Original Source | Check in Google Scholar | Full PDF (179.259 KB) | DOI: 10.14710/medstat.9.2.107-117

Abstract

Credit scoring models is an important tools in the credit granting process. These models measure the credit risk of a prospective client. This study aims to applied a discriminant model with mixed predictor variables (binary and continuous) for credit assesment. Implementation of the model use debitur characteristics data from a bank in Lampung Province which the used binary variables involve sex and marital status. Whereas, the continuous variables that was considered appropriate in the model are age, net income, and length of work. By using the data training, it was known that the misclassification of the model is 0.1970 and the misclassification of the testing data reach to 0.3753. Keywords: discriminant analysis, mixed variables, credit scoring