Articles

Found 39 Documents
Search

REGRESI SEMIPARAMETRIK SPLINE TRUNCATED DENGAN SOFTWARE R Utami, Tiani Wahyu; Prahutama, Alan
PROSIDING SEMINAR NASIONAL & INTERNASIONAL 2017: Prosiding Seminar Nasional Pendidikan, Sains dan Teknologi
Publisher : Universitas Muhammadiyah Semarang

Show Abstract | Download Original | Original Source | Check in Google Scholar | Full PDF (383.046 KB)

Abstract

Metode statisika sangat berperan penting dalam memprediksi maupunmenduga. Salah satu metode yang digunakan adalah dengan analisisregresi semiparametrik spline. Bentuk estimator Spline Truncated sangatdipengaruhi oleh nilai titik knots. Oleh karena itu, pemilihan titik knotoptimal mutlak diperlukan. Metode pemilihan titik knot dengan GCV.  Dua algoritma dan pemograman  untuk mendapatkan pemodelan regresisemiparametrik spline truncated dengan menggunakan software R yaitualgoritma dan program untuk menentukan Knot optimal berdasarkan metodeGCV, algoritma dan program untuk mengestimasi model regresisemiparametrik Spline Truncated. Keywords: Spline Truncated, GCV, Software R.
KLASIFIKASI INDEKS PEMBANGUNAN MANUSIA KABUPATEN/KOTA SE-INDONESIA DENGAN PENDEKATAN SMOOTH SUPPORT VECTOR MACHINE (SSVM) KERNEL RADIAL BASIS FUNCTION (RBF) Fauzi, Fatkhurokhman; Darsyah, Moh. Yamin; Utami, Tiani Wahyu
PROSIDING SEMINAR NASIONAL & INTERNASIONAL 2017: Prosiding Seminar Nasional Pendidikan, Sains dan Teknologi
Publisher : Universitas Muhammadiyah Semarang

Show Abstract | Download Original | Original Source | Check in Google Scholar | Full PDF (532.023 KB)

Abstract

Human Development Index (HDI) is a measure of human development achievementbased on basic components of quality of life. The human development index is low ifthe HDI is less than 60, moderate HDI between 60 to less than 70, high HDI between70 to less than 80, and equal to 80 and more than 80 are high. Smooth SupportVector Machine (SSVM) is a classification technique that is new. The algorithm usedis Radial Basis Function (RBF). The result of human development sperm using SSVMmethod with RBF kernel is 100%. With 41 districts / cities including low HDI. While332 districts / cities are included in medium HDI coverage, 134 districts / cities areincluded in the high HDI, and 12 districts / cities including HDI is very high. Keywords : Human Development Index, Smooth Support Vector Machine (SSVM), Radial Basis Function (RBF), accuracy, classification.
PEMODELAN REGRESI RIDGE PADA KASUS CURAH HUJAN DI KOTA SEMARANG Afham, Maulana; Nur, Indah Manfaati; Utami, Tiani Wahyu
PROSIDING SEMINAR NASIONAL & INTERNASIONAL 2017: Prosiding Seminar Nasional Pendidikan, Sains dan Teknologi
Publisher : Universitas Muhammadiyah Semarang

Show Abstract | Download Original | Original Source | Check in Google Scholar | Full PDF (983.537 KB)

Abstract

Rainfall is the amount of water that falls on the surface of the flat ground for acertain period measured in units of height (mm) above the horizontal surface. Theclassification of rainfall is divided into thick, medium, and light. Based on data of2016 Semarang city rainfall for 6 years experienced a significant decrease andincrease. With the data of rainfall Semarang city is very high potential for flooding.Semarang rainfall data tend to be unstable then it will cause problems in rainfalldata. Therefore it is necessary to solve the problem in rainfall data. The purpose ofthis study is to model and know the factors that affect rainfall in the city ofSemarang. The results of multiple regression found problems in Multicollinearity. An appropriate method for overcoming multiko in multiple regression is the ridgeregression. Regression of ridge to stabilize regression coefficient value of deviationof assumption in Multicolinearity. The result of the research to select the best model using the smallest MSE value which in the regex ridge model has MSE value 1.517smaller than the value of MSE in multiple regression of 1,519. While for variables that have significant effect on rainfall is wind speed, while variable temperature, humidity, solar irradiance have a significant influence but have weak effect on rainfall in Semarang city.Keywords: Rainfall, Ridge Regression and Multiple Regression
KERNEL NONPARAMETRIC REGRESSION FOR THE MODELIZING OF THE PRODUCTIVITY WETLAND PADDY Utami, Tiani Wahyu; Prihaswati, Martyana; Varima, Vega Zayu
PROSIDING SEMINAR NASIONAL & INTERNASIONAL 2018: PROCEEDING 1ST INSELIDEA INTERNATIONAL SEMINAR ON EDUCATION AND DEVELOPMENT OF ASIA (INseIDEA)
Publisher : Universitas Muhammadiyah Semarang

Show Abstract | Download Original | Original Source | Check in Google Scholar | Full PDF (376.949 KB)

Abstract

Nonparametric regression can be used when the relationship between the response variable and the predictor variables have an unknown pattern form the regression curve. One of the method that can be used to predictproductivity of the wetland paddy is a nonparametric regression kernel. In kernel regression, there are severaltypes of estimator that can be used to modelling productivity of wetland paddy in Central Java, one of which isNadaraya-Watson estimator. Variables used in the study of the productivity of rice as the response variable,while the predictor variables that harvested area, production and rainfall. Based on estimates indicate that thekernel nonparametric regression optimum bandwidth value 1.2 and GCV = 1.7577. The coefficient ofdetermination (R2) of 94.23% and MSE of 0.8560. Keywords: Kernel Nonparametric Regression, Productivity, Wetland Paddy
FAKTOR-FAKTOR DOMINAN YANG MEMPENGARUHI LAMA MENCARI PEKERJAAN DI SEMARANG MENGGUNAKAN ANALISIS REGRESI COX Sholiha, Anissatush; Wasono, Rochdi; Utami, Tiani Wahyu
PROSIDING SEMINAR NASIONAL & INTERNASIONAL 2017: Prosiding Seminar Nasional Pendidikan, Sains dan Teknologi
Publisher : Universitas Muhammadiyah Semarang

Show Abstract | Download Original | Original Source | Check in Google Scholar | Full PDF (521.936 KB)

Abstract

Discussion on the issue of unemployment is always associated with variousfactors that affect the length of time a person needs to get a job.One commonmethod used to determine these factors is to conduct survival analysis, amongwhich commonly used is Cox Proportional Hazard Regression.The purpose ofthis study was to identify the various factors of the time needed to be employeda university fresh graduate. The variables used consist of time to be employedas the dependent variable, while the independent variables are the educationalbackground, family income, job vacancies and the work aspiration. Coxregression can be the most appropriate method because the function andpurpose of this analysis is to predict exactly what factors make a person take acertain time to get his current job. The survival function and hazard functionpresent in the cox regression method allow the estimated time required by aperson until the person experiences an event (in which case the event is gettinga job). The results obtained from the analysis of each of these variables provedto have a significant effect on the length of time seeking employment of privateworkers in the city of Semarang.Keywords: Unemployed Length, Proportional Hazard Cox Regression, Survival.
GEOGRAPHICALLY WEIGHT REGRESSION APPLICATIONS FOR SPATIAL ANALYSIS OF INEQUALITY IN CENTRAL JAVA Janah, Lia Miftakhul; Nur, M. Saifudin; Utami, Tiani Wahyu
PROSIDING SEMINAR NASIONAL & INTERNASIONAL 2017: Proceeding 3rd ISET 2017 | International Seminar on Educational Technology 3rd 2017
Publisher : Universitas Muhammadiyah Semarang

Show Abstract | Download Original | Original Source | Check in Google Scholar | Full PDF (657.247 KB)

Abstract

Inequality is a state where there is an imbalance between each other. Inequality indicates the unevenness ofdevelopment that runs in an area. In Central Java, the problem of inequality among people still exists in daily life.Geographically Weight Regression method is a method that yields model parameter estimators that have localizedproperties at each point or location. In this study aims to modeling the inequality problem that occurred in CentralJava using Geographically Weight Regression method that has the nature of localization at the point. Data takenfrom Central Statistics Agency (BPS) 2015. Through Geographically Weight Regression method can beconcluded that with OLS method got 2 variables effect on imbalance wit h ? 10% is variable of HDI  (IPM) andPDRB. While the influential GWR method is the number of population and the amount of labor. While goodnessof fit test showed there is no difference between GWR model and OLS model or in other words there is no spatialeffect in the imbalance analysis in Central Java Province (0.4976 <0.1).Keywords: inequality, GWR,Spatial
ANALISIS SISTEM ANTRIAN MODEL MULTI PHASE-MULTI CHANNEL PADA SENTRA PELAYANAN KIOS 3 IN 1 BBPLK SEMARANG Rahayu, Ujiati Suci; Wasono, Rochdi; Utami, Tiani Wahyu
PROSIDING SEMINAR NASIONAL & INTERNASIONAL 2017: Prosiding Seminar Nasional Pendidikan, Sains dan Teknologi
Publisher : Universitas Muhammadiyah Semarang

Show Abstract | Download Original | Original Source | Check in Google Scholar | Full PDF (447.207 KB)

Abstract

The queue process is a process associated with the arrival of a customer at a service facility, then waited in a row (queue) when all services are busy, andleaving the place after getting the service. The queue process can happenanywhere, including in BBPLK Semarang. A wide variety of services such asregistration, competency testing, placement, making the rights of participantsand certificates. Therefore, it is necessary to study on the queuing system tooptimize service to customers. The purpose of this study was to determine thestatistical analysis deskriptive, making modeling a queue that services moreeffectively and efficiently and interpret the queuing models. The research in thispaper begins with a queuing system design kiosk 3 in 1 BBPLK Semarang.Then, the retrieval of data for each counter in the form of many arrivals anddepartures every 15 minutes. The collected data is then tested to determinewhether the data is distributed Poisson or not. Once known Poisson distributeddata, followed by determining a model queue at each phase and determine therate of arrivals and departures every service counter. The next step is to analyzethe size of the performance of each phase in the form of the average number ofcustomers in the system, the average number of customers in the queue, theaverage length of customer in the system, and the average length of customer inthe queue.  Keywords : queue, model multi phase-multi channel,  poisson, eksponensial
ESTIMASI KURVA REGRESI SEMIPARAMETRIK PADA DATA LONGITUDINAL BERDASARKAN ESTIMATOR POLINOMIAL LOKAL Utami, Tiani Wahyu
Jurnal Statistika Vol 1, No 1 (2013): Jurnal Statistika
Publisher : Jurnal Statistika

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

Abstract

Diberikan model regresi semiparametrik untuk data longitudinal , dengan komponen parametrik  dan komponen nonparametrik yang didekati dengan Polinomial Lokal. Estimator Polinomial Lokal diperoleh dengan metode WLS (Weighted Least Square). Estimasi model regresi semiparametrik pada data longitudinal adalah , dengan , Estimator Polinomial Lokal sangat tergantung pada bandwidth (h) optimal. Penentuan bandwidth optimal dengan menggunakan metode GCV (Generalized Cross Validation). Selanjutnya model regresi semiparametrik Polinomial Lokal pada data longitudinal diaplikasikan untuk memodelkan hubungan pengaruh antara kadar trombosit penderita Demam Berdarah Dengue terhadap kadar hematrokit dan waktu pemeriksaan, dimana kadar hematrokit sebagai komponen nonparametrik dan waktu pemeriksaan sebagai komponen parametrik. Hasil estimasi menujukkan bahwa waktu pemeriksaan berpola kuadratik untuk setiap subjek sedangkan kadar hematrokit pada subjek 1 mengikuti pola polinomial lokal berorde 1, sedangkan kadar hematrokit pada subjek 2 mengikuti pola polinomial lokal berorde 4 dan pada subjek 3 kadar hematrokit memgikuti pola polinomial lokal berorde 4. Model ini mempunyai nilai MSE sebesar 146.7636 dan koefisien determinasi R2 = 93,9249 %.   Kata kunci : Data Longitudinal, Estimator Polinomial Lokal, GCV, Regresi Semiparametrik, WLS.  
ANALISIS REGRESI BINOMIAL NEGATIF UNTUK MENGATASI OVERDISPERSION REGRESI POISSON PADA KASUS DEMAM BERDARAH DENGUE Utami, Tiani Wahyu
Jurnal Statistika Vol 1, No 2 (2013): Jurnal Statistika
Publisher : Jurnal Statistika

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

Abstract

Dalam analisis regresi Poisson, variabel respon (Y) harus memenuhi asumsiequidispersion (nilai varians sama dengan mean). Namun, dalam data riil seringkaliterjadi overdispersion (nilai varians lebih besar dari mean). Salah satu cara untukmengatasinya yaitu dengan mengganti asumsi distribusi Poisson dengan distribusiBinomial Negatif. Tujuan dari artikel ini adalah mendapatkan pola hubungan terbaikdalam analisis regresi Binomial Negatif untuk mengatasi overdispersion regresi Poisson  Kasus Demam Berdarah Dengue pada Balita Menurut Kabupaten/Kota  di Propinsi  Jawa Timur. Berdasarkan persamaan model regresi Binomial Negatif yang diperoleh dapat dijelaskan bahwa dengan semakin bertambahnya presentase tenaga medis di sarana pelayanan kesehatan dan presentase rumah tangga yang memiliki rumah sehat , maka akan menurunkan jumlah penderita DBD pada balita di Propinsi Jawa Timur. Kata Kunci :  Regresi Poisson, Equidispersion, Overdispersion, Generalized LinierModel (GLM), Regresi Binomial Negatif.
PEMODELAN ANGKA KEMATIAN BAYI DENGAN PENDEKATAN REGRESI NONPARAMETRIK SPLINE TRUNCATED Utami, Tiani Wahyu
PROSIDING SEMINAR NASIONAL & INTERNASIONAL 2018: SEMINAR NASIONAL PENDIDIKAN SAINS DAN TEKNOLOGI
Publisher : Universitas Muhammadiyah Semarang

Show Abstract | Download Original | Original Source | Check in Google Scholar | Full PDF (504.41 KB)

Abstract

Kematian bayi merupakan salah satu indikator dalam menentukan derajat kesehatan. Apabila suatu daerah memiliki kematian bayi yang tinggi maka dapat dikatakan tingkat kesehatan anak  pada daerah tersebut rendah. Angka kematian bayi juga mampu menggambarkan keadaan sosial di masyarakat.Tujuan dari penelitian ini adalah untuk memodelkan antara variabel prediktor dengan variabel respon. Variabel yang diduga adalah (Y) Angka Kematian Bayi (AKB), persentasi bayi yang diberi asi ekslusif (X1) dan persentase persalinan dengan tenaga medis (X2). Metode ini digunakan dalam penelitian ini adalah Regresi Spline Truncated, model ini cenderung mencari sendiri estimasi data. Dalam metode ini terdapat titik knot, yaitu titik yang menunjukan perubahan data. Pemilihan titik knot optimum dilakukan dengan cara memilih nilai Generalized Cross Validation (GCV) yang minimum. Niliai GCV terkecil sebesar  5578.896 dengan R2 sebesar 86,551%.Keywords : Kematian Bayi, Regresi Spline Truncated, GCV.